24.06.2014 Views

thèse pour rapporteur - Station Biologique de Roscoff

thèse pour rapporteur - Station Biologique de Roscoff

thèse pour rapporteur - Station Biologique de Roscoff

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

THESE DE DOCTORAT DE<br />

L’UNIVERSITE PIERRE ET MARIE CURIE<br />

Spécialité<br />

Diversité du Vivant<br />

Présentée par<br />

Mlle Sophie PLOUVIEZ<br />

Pour obtenir le gra<strong>de</strong> <strong>de</strong><br />

DOCTEUR <strong>de</strong> l’UNIVERSITÉ PIERRE ET MARIE CURIE<br />

Phylogéographie comparée <strong>de</strong>s espèces<br />

hydrothermales <strong>de</strong> la dorsale du Pacifique oriental<br />

soutenue le 11 décembre 2009<br />

<strong>de</strong>vant le jury composé <strong>de</strong> :<br />

M. Dominique HIGUET, Professeur, Université Pierre et Marie Curie (Prési<strong>de</strong>nt du jury)<br />

Mme. Sophie ARNAUD-HAOND, Chargée <strong>de</strong> recherche, Ifremer (Rapporteur)<br />

M. François BONHOMME, Directeur <strong>de</strong> Recherche, CNRS, Montpellier (Rapporteur<br />

Mme. Sarah SAMADI, Chargée <strong>de</strong> recherche, IRD<br />

(Examinateur)<br />

M. Pierre CHEVALDONNE, Chargé <strong>de</strong> recherche, Endoume (Examinateur)<br />

M. Didier JOLLIVET, Chargé <strong>de</strong> recherche, CNRS, <strong>Roscoff</strong> (Directeur <strong>de</strong> thèse)<br />

M. François LALLIER, Professeur, Université Pierre et Marie Curie (Directeur <strong>de</strong> thèse)


Table <strong>de</strong>s matières


TABLE DES MATIÈRES<br />

CHAPITRE 1 : INTRODUCTION GENERALE A LA PHYLOGEOGRAPHIE ............................................1<br />

1. BIOGEOGRAPHIE ET PHYLOGEOGRAPHIE .....................................................................................................1<br />

2. MODELES DE DISPERSION EN POPULATIONS GEOGRAPHIQUEMENT SUBDIVISEES......................................3<br />

3. APPROCHE DE LA PHYLOGEOGRAPHIE PAR LA THEORIE DE LA COALESCENCE ..........................................9<br />

3.1. Principe <strong>de</strong> la théorie <strong>de</strong> la coalescence............................................................................................9<br />

3.2. Écart à la neutralité : impact <strong>de</strong> l’histoire démographique et <strong>de</strong> la sélection naturelle sur le<br />

coalescent.......................................................................................................................................................13<br />

3.3. Structure du coalescent au regard <strong>de</strong> la géographie.......................................................................16<br />

3.4. Coalescence <strong>de</strong> gènes versus séparation <strong>de</strong>s lignées intra-spécifiques .........................................20<br />

4. RELACHEMENT D’UN ISOLEMENT GEOGRAPHIQUE....................................................................................22<br />

5. BIOGEOGRAPHIE EN MILIEU HYDROTHERMAL PROFOND...........................................................................24<br />

6. DORSALE DU PACIFIQUE ORIENTAL (EPR) ................................................................................................30<br />

6.1. Vitesses d’accrétion et géomorphologie du rift ...............................................................................30<br />

6.2. Définition et principe d’une faille transformant et d’une microplaque .........................................31<br />

6.3. Histoire <strong>de</strong> la formation <strong>de</strong> la dorsale EPR.....................................................................................32<br />

6.3.1. La microplaque Bauer: 10°S-18°S ............................................................................................................. 34<br />

6.3.2. La paléoplaque Mathematician: 12°N-17°N ............................................................................................. 34<br />

6.3.3. Capture <strong>de</strong> la microplaque Bauer et formation <strong>de</strong>s microplaques <strong>de</strong> Pâques.......................................... 35<br />

6.4. Potentielles barrières à la dispersion...............................................................................................36<br />

6.4.1. Failles transformantes ................................................................................................................................. 36<br />

6.4.2. Courants océaniques profonds.................................................................................................................... 37<br />

7. STRUCTURE GENETIQUE DES ESPECES DE LA DORSALE DU PACIFIQUE ORIENTAL...................................38<br />

7.1. Quels modèles <strong>de</strong> dispersion <strong>pour</strong> les espèces <strong>de</strong> l’EPR ? .............................................................38<br />

7.2. Existence <strong>de</strong> barrières génétiques chez les espèces <strong>de</strong> l’EPR .......................................................40<br />

8. OBJECTIFS DE LA THESE..............................................................................................................................42<br />

CHAPITRE 2 : METHODES ET BIAIS METHODOLOGIQUES .................................................................. 43<br />

1. STRATEGIE D’ECHANTILLONNAGE .............................................................................................................43<br />

2. BANQUES D’ADN COMPLEMENTAIRES .....................................................................................................46<br />

3. ACQUISITION DES SEQUENCES....................................................................................................................47<br />

3.1. Séquences du gène mitochondrial Cytochrome Oxydase I..............................................................47<br />

3.2. Séquences <strong>de</strong>s gènes nucléaires : Marquage-Recapture (MR) <strong>de</strong> séquences................................48<br />

4. TRAITEMENT DES SEQUENCES ....................................................................................................................50<br />

4.1. Alignement et correction <strong>de</strong>s séquences liées aux erreurs <strong>de</strong> lecture du séquenceur...................50<br />

4.2. Recombinaison in vitro détectée via la métho<strong>de</strong> MR.......................................................................51<br />

4.3. Correction du jeu <strong>de</strong> données : suppression <strong>de</strong>s singletons artéfactuels engendrant un<br />

multiallélisme.................................................................................................................................................52<br />

5. TRAITEMENT DE LA RECOMBINAISON ........................................................................................................54


5.1. Biais liés à la recombinaison dans les analyses <strong>de</strong> coalescence....................................................54<br />

5.2. Détection et suppression <strong>de</strong>s séquences recombinantes .................................................................55<br />

6. DETECTION DE GENES CO-AMPLIFIES (POTENTIELLEMENT PARALOGUES)...............................................55<br />

6.1. Bathymodiolus thermophilus.............................................................................................................55<br />

6.1.1. Lysozyme .....................................................................................................................................................55<br />

6.1.2. Sulfotransférase............................................................................................................................................56<br />

6.2. Lepetodrilus elevatus.........................................................................................................................57<br />

6.2.1. Lysozyme .....................................................................................................................................................58<br />

6.2.2. Veliger Digestive Gland 3 (VDG3)............................................................................................................59<br />

6.2.3. Textilinine ....................................................................................................................................................62<br />

7. RECHERCHE D’UNE DIVERGENCE PARTAGEE ENTRE PLUSIEURS ESPECES : MSBAYES............................63<br />

8. MIGRATION A TRAVERS UNE BARRIERE AUX FLUX DE GENES : MIGRATE-N............................................66<br />

9. MODELE D’ « ISOLEMENT AVEC MIGRATION ».........................................................................................68<br />

10. ÉCART A LA NEUTRALITE : SELECTION VERSUS DEMOGRAPHIE .............................................................69<br />

10.1. Approche multilocus visant à détecter un ou plusieurs locus sous sélection positive : test <strong>de</strong><br />

Hudson Kreitman Aguadé.............................................................................................................................69<br />

10.2. Approche substitutions synonymes / non-synonymes en vu <strong>de</strong> la détection <strong>de</strong> sélection positive<br />

versus négative : test <strong>de</strong> MacDonald-Kreitman ..........................................................................................70<br />

10.3. Tests <strong>de</strong> polymorphisme visant à détecter un écart à la neutralité : tests <strong>de</strong> Tajima et Fu &<br />

Li….................................................................................................................................................................71<br />

10.4. Dater une expansion démographique via le logiciel Fluctuate....................................................72<br />

10.5. Proposition d’un nouveau test statistique visant à détecter <strong>de</strong>s locus sous sélection<br />

diversifiante entre régions géographiques ..................................................................................................73<br />

10.6. Discrimination entre balayage sélectif et goulôt d’étranglement via une approche <strong>de</strong><br />

vraisemblance sur données multilocus : le logiciel Sweep_bott ...............................................................74<br />

CHAPITRE 3 : PHYLOGEOGRAPHIE COMPAREE DES ESPECES HYDROTHERMALES DE<br />

LA DORSALE DU PACIFIQUE ORIENTAL A L’AIDE DU GENE MITOCHONDRIAL DE LA<br />

CYTOCHROME OXYDASE I.................................................................................................................................77<br />

1. INTRODUCTION A L’APPROCHE DE PHYLOGEOGRAPHIE COMPAREE DES ESPECES ...................................77<br />

2. RESUME DE L’ARTICLE DE PHYLOGEOGRAPHIE COMPAREE ENTRE ESPECES............................................81<br />

3. ARTICLE : PHYLOGEOGRAPHIE COMPAREE ENTRE ESPECES SUR LE GENE MTCOI ..................................87<br />

4. INTERETS ET LIMITES DE L’APPROCHE DE PHYLOGEOGRAPHIE COMPAREE............................................102<br />

4.1. Apports <strong>de</strong> la phylogéographie comparée entre espèces dans l’étu<strong>de</strong> <strong>de</strong>s processus <strong>de</strong><br />

vicariance ....................................................................................................................................................102<br />

4.2. Apports <strong>de</strong> la phylogéographie comparée dans l’étu<strong>de</strong> <strong>de</strong> l’histoire démographique <strong>de</strong>s<br />

populations ..................................................................................................................................................104<br />

4.3. Limites <strong>de</strong> la phylogéographie comparée monogène ....................................................................105<br />

CHAPITRE 4 : ETUDE D’UNE BARRIERE SEMI-PERMEABLE A LA DISPERSION PAR UNE<br />

APPROCHE MULTI-LOCUS SUR PLUSIEURS ESPECES ......................................................................... 107<br />

1. INTRODUCTION A L’ETUDE D’UNE BARRIERE A LA DISPERSION PAR UNE APPROCHE MULTI-LOCUS.....107


2. TAXONOMIE ET TRAITS D’HISTOIRE DE VIE DES ESPECES CIBLES .......................................................... 111<br />

2.1. A. pompejana................................................................................................................................... 111<br />

2.2. B. thermophilus ............................................................................................................................... 113<br />

2.3. L. elevatus........................................................................................................................................ 116<br />

3. MATERIELS ET METHODES DES ESPECES CIBLES..................................................................................... 118<br />

4. ANALYSES ET RESULTATS SUR LES ESPECES CIBLES .............................................................................. 118<br />

4.1. A.pompejana.................................................................................................................................... 118<br />

4.1.1. Résumé <strong>de</strong>s résultats ................................................................................................................................. 118<br />

4.1.2. Article en préparation................................................................................................................................125<br />

4.2. B. thermophilus ............................................................................................................................... 170<br />

4.2.1. Résumé <strong>de</strong>s résultats ................................................................................................................................. 170<br />

4.2.2. Article en préparation................................................................................................................................175<br />

4.3. L. elevatus........................................................................................................................................ 216<br />

4.3.1. Introduction................................................................................................................................................216<br />

4.3.2. Gène mitochondrial COI...........................................................................................................................217<br />

4.3.3. Analyse <strong>de</strong>s séquences nucléaires : hybridation et barrière(s) au(x) flux <strong>de</strong> gènes ..............................220<br />

4.3.4. Y a-t-il eu une expansion démographique au Sud <strong>de</strong> l’EPR ?................................................................235<br />

5. DISCUSSION GENERALE DU CHAPITRE 4 : COMPARAISON DES RESULTATS OBTENUS ENTRE LES ESPECES<br />

CIBLES............................................................................................................................................................... 239<br />

CHAPITRE 5 : CONCLUSION, DISCUSSION GENERALE ET PERSPECTIVES ................................. 245<br />

1. L’EPR : UNE OU DEUX PROVINCES BIOGEOGRAPHIQUES ?.................................................................... 246<br />

2. CONNEXION ENTRE EPR ET LES AUTRES PROVINCES BIOGEOGRAPHIQUES .......................................... 249<br />

3. BARRIERES SEMI-PERMEABLES : ROLE DES FAILLES TRANSFORMANTES ET DES COURANTS<br />

DIVERGENTS ? .................................................................................................................................................. 250<br />

4. PERSPECTIVES DE LA THESE .................................................................................................................... 251<br />

4.1. Évolution <strong>de</strong>s espèces hydrothermales .......................................................................................... 251<br />

4.2. Évolution <strong>de</strong>s gènes......................................................................................................................... 253<br />

4.2.1. Le gène mtCOI ..........................................................................................................................................253<br />

4.2.2. Phosphoglucomutase d’A. pompejana ..................................................................................................... 255<br />

5. ANNEXE DU CHAPITRE 5 .......................................................................................................................... 256<br />

REFERENCES BIBLIOGRAPHIQUES ............................................................................................................. 299<br />

ANNEXE 1 : MARQUAGE RECAPTURE DE SEQUENCES........................................................................ 323<br />

ANNEXE 2 : LOGICIEL RABOUTER SEQUENCE ....................................................................................... 329<br />

ANNEXE 3 : CONDITIONS DE PCR.................................................................................................................. 341<br />

ANNEXE 4 : CODE GENETIQUE MITOCHONDRIAL DES ALVINELLIDAE...................................... 345<br />

ANNEXE 5 : ARTICLE RELATION FEMELLES/JUVENILES CHEZ BRANCHIPOLYNOE<br />

SEEPENSIS............................................................................................................................................................... 349


Introduction générale<br />

Chapitre 1 : Introduction générale à la<br />

phylogéographie<br />

1. Biogéographie et phylogéographie<br />

Subdiviser le mon<strong>de</strong> vivant en multiples unités taxonomiques a permis aux<br />

scientifiques <strong>de</strong> mieux en appréhen<strong>de</strong>r sa diversité. Dès lors, <strong>de</strong>s patrons communs <strong>de</strong><br />

distribution géographique ont pu être mis en évi<strong>de</strong>nce entre taxons, posant ainsi la question<br />

<strong>de</strong>s raisons historiques ayant conduit à cette répartition particulière. Repérer ces<br />

distributions communes à l’échelle <strong>de</strong>s peuplements et i<strong>de</strong>ntifier les processus historiques,<br />

environnementaux et biologiques, qui les ont façonnées, tel est l’objectif <strong>de</strong> la<br />

biogéographie (Myers & Giller, 1988). L’étu<strong>de</strong> <strong>de</strong>s compositions faunistiques et<br />

floristiques a déjà permis l’i<strong>de</strong>ntification <strong>de</strong> nombreuses provinces biogéographiques, aussi<br />

bien en milieu terrestre (e.g. Cox 2001) qu’en milieu aquatique (e.g. Cox & Moore 2000).<br />

Si les barrières biogéographiques terrestres sont, du moins au premier abord, souvent<br />

relativement faciles à établir (correspondant à la séparation <strong>de</strong>s continents par les océans<br />

ou à l’existence <strong>de</strong> montagnes ou <strong>de</strong> déserts), l’interconnexion <strong>de</strong>s océans ainsi que la<br />

difficulté d’accessibilité <strong>de</strong> ce milieu rend plus complexe la visualisation <strong>de</strong> barrières à la<br />

dispersion et donc la définition <strong>de</strong> provinces biogéographiques marines. L’adaptation <strong>de</strong>s<br />

organismes à un environnement donné est également à prendre en considération lors <strong>de</strong> la<br />

définition <strong>de</strong> provinces biogéographiques, rendant la tâche relativement complexe (e.g.<br />

continuum <strong>de</strong> températures). La profon<strong>de</strong>ur <strong>de</strong>s océans peut ainsi jouer un rôle<br />

fondamental dans l’étu<strong>de</strong> <strong>de</strong> la biogéographie en limitant l’impact <strong>de</strong>s changements<br />

climatiques sur la distribution <strong>de</strong>s espèces (Cox & Moore 2000). En effet, en cas <strong>de</strong><br />

changement <strong>de</strong> température, par exemple, les espèces peuvent atteindre leur preferendum<br />

thermique par migration latitudinale, comme en milieu terrestre, mais également par<br />

migration bathymétrique (en milieu terrestre, l’altitu<strong>de</strong> peut également jouer ce rôle mais<br />

les points élevés restent relativement localisés).<br />

1


Introduction générale<br />

La subdivision d’une région ancestrale mère en <strong>de</strong>ux provinces biogéographiques filles<br />

peut être la conséquence <strong>de</strong> mécanismes divers. De nombreuses hypothèses ont ainsi été<br />

proposées <strong>pour</strong> expliquer ces subdivisions :<br />

• <strong>de</strong>s changements paléo-environnementaux liés aux mouvements tectoniques <strong>de</strong>s<br />

plaques (théorie paléogéographique, Emsley, 1965) ou aux oscillations<br />

climatiques (e.g. théorie <strong>de</strong>s refuges permettant aux zones géographiques les<br />

plus stables d'abriter certains taxons lors <strong>de</strong> contextes climatiques défavorables,<br />

Haffer, 1969)<br />

• d’autres barrières physiques à la dispersion qui fragmentent l’habitat tels que<br />

<strong>de</strong>s courants océaniques, un fleuve, un désert…<br />

• <strong>de</strong>s gradients environnementaux (Endler 1982) qui, par sélection naturelle,<br />

peuvent engendrer <strong>de</strong>s gradients <strong>de</strong> distribution <strong>de</strong>s taxons<br />

Quelle que soit l’hypothèse, les capacités <strong>de</strong> dispersion et d’adaptation <strong>de</strong>s espèces<br />

jouent un rôle fondamental dans leur répartition géographique et permettent <strong>de</strong> coloniser<br />

<strong>de</strong> nouveaux territoires, <strong>de</strong> survivre aux conditions locales <strong>de</strong> nouveaux habitats et s’y<br />

maintenir. Ainsi, définir une province biogéographique implique que <strong>de</strong> multiples espèces<br />

ne présentent pas les capacités <strong>de</strong> dispersion et/ou une valeur sélective suffisantes <strong>pour</strong><br />

atteindre la province biogéographique voisine et s’y établir. Néanmoins, certaines espèces,<br />

<strong>de</strong> par leur nature plus ubiquiste, sont capables <strong>de</strong> traverser une barrière biogéographique.<br />

Cela ne signifie pas <strong>pour</strong> autant qu’elles ne sont pas affectées par la barrière<br />

biogéographique, certains individus pouvant être plus dispersifs ou mieux adaptés aux<br />

contraintes <strong>de</strong> l’environnement. Pour évaluer l’impact intraspécifique <strong>de</strong> ces contraintes, il<br />

est donc nécessaire d'examiner le polymorphisme <strong>de</strong> ces espèces et d’apprécier les<br />

différences potentielles pouvant exister entre populations situées <strong>de</strong> part et d’autre <strong>de</strong> la<br />

barrière biogéographique (ségrégation allélique <strong>pour</strong> un caractère donné). L’utilisation<br />

d’outils moléculaires permet <strong>de</strong> comparer un nombre plus important <strong>de</strong> caractères<br />

considérés neutres (e.g. certains sites nucléotidiques), qu’une approche morphologique, par<br />

exemple, et maximise la probabilité <strong>de</strong> détecter cet impact potentiel. Ainsi, une analyse <strong>de</strong><br />

la structure génétique <strong>de</strong> ces populations peut permettre <strong>de</strong> détecter une concordance entre<br />

barrière génétique et barrière biogéographique, et ainsi indiquer que cette <strong>de</strong>rnière a<br />

également eu un effet d’isolement sur l’espèce étudiée. Par exemple, la barrière<br />

biogéographique <strong>de</strong> Cap Canaveral, qui sépare les faunes <strong>de</strong> l’Atlantique et du Golfe du<br />

Mexique, a également engendré <strong>de</strong> la différenciation génétique intraspécifique <strong>de</strong> part et<br />

2


Introduction générale<br />

d’autre <strong>de</strong> cette barrière (Avise 1992). De même, les populations <strong>de</strong> multiples taxons<br />

distribués <strong>de</strong> part et d’autre <strong>de</strong> l’archipel indo-australien présentent une divergence<br />

génétique accrue entre les océans Indien et Pacifique occi<strong>de</strong>ntal (e.g. crabe <strong>de</strong>s cocotiers<br />

Birgus latro : Lavery et al, 1995; crevette tigrée Penaeus monodon : Duda & Palumbi,<br />

1999). Néanmoins, les barrières biogéographiques ne correspon<strong>de</strong>nt pas nécessairement à<br />

<strong>de</strong>s subdivisions génétiques (évolution en allopatrie). La divergence intraspécifique <strong>de</strong> part<br />

et d’autre <strong>de</strong> la barrière biogéographique <strong>de</strong> Point Conception (séparant la Californie <strong>de</strong><br />

l’Oregon) est extrêmement faible, indiquant que cette barrière biogéographique n’aurait<br />

pas réellement joué le rôle <strong>de</strong> barrière intraspécifique en limitant les flux <strong>de</strong> gènes entre<br />

populations (Burton 1998). Lorsqu’elle existe, la relation entre barrière génétique et<br />

barrière biogéographique n’est donc pas toujours simple à mettre à en évi<strong>de</strong>nce, d’autant<br />

plus que la localisation d’une barrière à la dispersion peut varier selon les espèces,<br />

engendrant <strong>de</strong>s zones <strong>de</strong> transition plus ou moins larges entre provinces biogéographiques<br />

(e.g. zone <strong>de</strong> transition le long <strong>de</strong>s côtes chiliennes, Thiel et al 2007).<br />

La phylogéographie a <strong>pour</strong> objectif <strong>de</strong> comprendre les processus qui gouvernent la<br />

distribution géographique <strong>de</strong>s lignées intraspécifiques (Avise et al 1987) par l’étu<strong>de</strong> <strong>de</strong><br />

l’histoire évolutive <strong>de</strong> leurs gènes. Le processus <strong>de</strong> subdivision phylogéographique le plus<br />

couramment admis peut ainsi se décliner en plusieurs étapes : (1) un isolement<br />

géographique en allopatrie dépendant notamment du modèle <strong>de</strong> dispersion <strong>de</strong> l’espèce<br />

considérée et <strong>de</strong> la force <strong>de</strong> la barrière rencontrée ; pouvant engendrer, accumulation <strong>de</strong><br />

mutations <strong>de</strong> part et d’autre <strong>de</strong> la barrière et dérive génétique, et/ou maladaptation<br />

géographique, (2) un isolement génétique correspondant à cette subdivision géographique.<br />

2. Modèles <strong>de</strong> dispersion en populations géographiquement subdivisées<br />

Si le modèle panmictique correspond à un appariement sans contrainte <strong>de</strong>s<br />

individus au sein <strong>de</strong> la ou <strong>de</strong>s populations considérées (indice <strong>de</strong> différenciation génétique<br />

Fst = 0), à l’inverse, certaines populations subdivisées géographiquement peuvent<br />

n’échanger aucun migrant et conduire à une augmentation significative <strong>de</strong> la<br />

différenciation génétique par dérive (Hartl & Clark 2007). De nombreux modèles<br />

intermédiaires impliquant un échange <strong>de</strong> migrants sous contraintes ont alors été décrits,<br />

3


Introduction générale<br />

selon un <strong>de</strong>gré plus ou moins variable <strong>de</strong> connectivité <strong>de</strong>s populations géographiquement<br />

subdivisées.<br />

Wright (1931, 1937) intègre la connectivité dans son modèle en îles (Fig. 1.1.A) en<br />

y ajoutant un paramètre <strong>de</strong> migration entre les populations étudiées. Ainsi, toutes les<br />

populations sont supposées présenter le même taux <strong>de</strong> migration m avec l’ensemble <strong>de</strong>s<br />

autres populations. La différenciation entre les populations (mesurée par le Fst) peut ainsi<br />

permettre d’estimer la fraction d’individus (N x m) échangés entre populations (migrants)<br />

par la relation Fst = 1/(1+4Nm) sous l’hypothèse d’un état d’équilibre entre la migration et<br />

la dérive à chaque génération (populations démographiquement stables au cours du temps).<br />

Néanmoins, supposer que <strong>de</strong>ux populations proches échangent entre elles autant <strong>de</strong><br />

migrants que <strong>de</strong>ux populations éloignées ne semble pas très réaliste. C’est <strong>pour</strong>quoi,<br />

d’autres modèles intègrent la notion <strong>de</strong> proximité géographique entre populations.<br />

Le modèle en pas japonais (ou stepping stone, Kimura 1953, Fig. 1.1.B) considère<br />

que la dispersion ne s’effectue qu’entre populations voisines selon un même taux <strong>de</strong><br />

migration m (celui-ci étant nul entre populations non adjacentes). Le modèle d’isolement<br />

par la distance (Wright 1943, Malécot 1950, Fig. 1.1.C) présente, quant à lui, un taux <strong>de</strong><br />

migration proportionnel à la distance entre individus considérés. Ainsi, ce modèle permet<br />

la connectivité entre individus provenant <strong>de</strong> populations proches mais également, dans une<br />

moindre mesure, entre populations plus distantes (Wright 1943). Il suppose néanmoins que<br />

seule la distance est responsable du coefficient d’apparentement entre individus et<br />

n’intègre pas la présence potentielle d’une rupture abrupte <strong>de</strong> la migration par une barrière<br />

à la dispersion.<br />

Le modèle avec matrice <strong>de</strong> migration libre (Bodmer & Cavalli-Sforza 1968, Fig.<br />

1.1.D) permet <strong>de</strong> fixer un taux <strong>de</strong> migration différent entre chaque paire <strong>de</strong> populations,<br />

avec <strong>de</strong>s flux pouvant être asymétriques. Ainsi, selon les paramètres <strong>de</strong> migration choisis,<br />

ce modèle peut prendre en compte à la fois l’isolement par la distance mais également <strong>de</strong><br />

possibles barrières à la dispersion (migration réduite entre <strong>de</strong>ux groupes <strong>de</strong> populations).<br />

Les modèles en îles, en pas japonais et d’isolement par la distance, constituent <strong>de</strong>s cas<br />

particuliers du modèle avec matrice <strong>de</strong> migration libre. Néanmoins, intégrant les avantages<br />

<strong>de</strong> tous les modèles précé<strong>de</strong>nts, ce <strong>de</strong>rnier nécessite l’estimation d’un nombre important <strong>de</strong><br />

4


Introduction générale<br />

paramètres (n(n-1) paramètres <strong>de</strong> migration avec n correspondant au nombre <strong>de</strong><br />

populations). C’est <strong>pour</strong>quoi il n’est que rarement utilisé.<br />

A. Modèle en île B. Modèle en pas japonais<br />

C. Modèle d’isolement<br />

par la distance<br />

D. Modèle avec matrice<br />

<strong>de</strong> migration libre<br />

Fig. 1.1 Modèles <strong>de</strong> dispersion en populations géographiquement subdivisées.<br />

Les taux <strong>de</strong> migration sont proportionnels à l’épaisseur <strong>de</strong>s flèches.<br />

5


Introduction générale<br />

Selon les capacités <strong>de</strong> dispersion d’une espèce, un ajustement à un modèle <strong>de</strong> populations<br />

peut permettre <strong>de</strong> mieux appréhen<strong>de</strong>r son évolution dans le temps et l’espace. Ainsi, une<br />

espèce à forte capacité <strong>de</strong> dispersion sera moins soumise à l’isolement par la distance<br />

qu’une espèce à capacité <strong>de</strong> dispersion réduite et tendra à suivre un modèle en îles. Les<br />

espèces à capacité <strong>de</strong> migration très faible seront, quant à elles, plus proches d’un modèle<br />

d’isolement par la distance ou d’un modèle en pas japonais (suivant la fragmentation <strong>de</strong><br />

l’habitat). L’isolement par la distance sera d’autant plus marqué que les capacités <strong>de</strong><br />

dispersion <strong>de</strong>s espèces seront faibles (Palumbi 2003, Fig 1.2).<br />

Potentiel <strong>de</strong> dispersion :<br />

1 site par génération<br />

2 sites par génération<br />

3 sites par génération<br />

4 sites par génération<br />

Gst<br />

Distance (en nombre <strong>de</strong> sites) entre chacune <strong>de</strong>s 10 populations échantillonnées<br />

Fig. 1.2 Variations <strong>de</strong> la force <strong>de</strong> l’isolement par la distance avec la capacité <strong>de</strong><br />

dispersion larvaire moyenne (en nombre <strong>de</strong> sites atteints par génération, la distance<br />

entre sites étant constante). Adaptée <strong>de</strong> Palumbi (2003).<br />

Valeurs du Gst (analogue du Fst et mesurant la proportion <strong>de</strong> variabilité génétique qui<br />

est distribuée entre populations) simulées sur 10 locus bi-alleliques<br />

Si ces modèles permettent <strong>de</strong> décrire la connectivité entre populations à un temps<br />

donné, ils n’intègrent cependant pas la nature dynamique <strong>de</strong>s populations dans le temps.<br />

Les modèles <strong>de</strong> métapopulation (modèle initial proposé par Levin 1970 puis différents<br />

modèles dérivés revus dans Harrison & Hastings 1996 ; Fig 1.3) ont été introduits afin <strong>de</strong><br />

tenir compte <strong>de</strong> la variabilité dynamique locale <strong>de</strong>s populations en y intégrant la possibilité<br />

d’extinction et <strong>de</strong> recolonisation <strong>de</strong> ces populations par <strong>de</strong> nouveaux migrants (notion <strong>de</strong><br />

colonisateurs).<br />

6


Introduction générale<br />

A<br />

B<br />

C<br />

D<br />

E<br />

Fig. 1.3 Différents types <strong>de</strong> métapopulation. (A) type « classique » (Levins) ; (B) type<br />

« source-puits » ; (C) type « population fragmentée » ; (D) métapopulation en déséquilibre ;<br />

(E) cas intermédiaire combinant (A) et (D).<br />

Symboles remplis : habitats occupés ; symboles vi<strong>de</strong>s : habitats vierges ; flèches :<br />

dispersion; lignes noires entourant les symboles remplis : limite <strong>de</strong> distribution <strong>de</strong>s<br />

populations locales. D’après Harrison & Hastings (1996).<br />

Deux modèles <strong>de</strong> recolonisation ont été proposés (Slatkin 1977) : le modèle<br />

« migrant » (recolonisation d’un nouveau site à partir <strong>de</strong> colonisateurs provenant <strong>de</strong><br />

l’ensemble <strong>de</strong>s autres populations <strong>de</strong> la métapopulation) et le modèle « propagule »<br />

(recolonisation d’un site à partir <strong>de</strong> colonisateurs provenant uniquement <strong>de</strong>s populations<br />

voisines). Il a été montré que le taux d’extinction <strong>de</strong>s populations locales et le modèle <strong>de</strong><br />

recolonisation influencent fortement la structure génétique <strong>de</strong> la métapopulation lorsque<br />

ces modèles sont comparés aux modèles <strong>de</strong> référence sans extinction (Wa<strong>de</strong> & McCauley<br />

1988, Whitlock & McCauley 1990, Pannell & Charlesworth 2000).<br />

Dans le modèle « propagule », la recolonisation ne pouvant s’effectuer qu’à partir <strong>de</strong>s<br />

populations proches, les populations recolonisées présenteront alors une diversité<br />

génétique réduite (notion d’effet fondateur : cf. Pannell & Charlesworth 2000). La<br />

7


Introduction générale<br />

différenciation génétique (Fst) entre les populations augmentera avec le taux d’extinction<br />

(Wa<strong>de</strong> & McCauley 1988, Whitlock & McCauley 1990).<br />

Dans le modèle « migrant », la recolonisation s’effectue à partir d’individus provenant <strong>de</strong><br />

l’ensemble <strong>de</strong>s autres populations. Si le nombre <strong>de</strong> colonisateurs est suffisamment grand (><br />

2 fois le nombre <strong>de</strong> migrants échangés entre populations déjà établies), alors la population<br />

recolonisée présentera une diversité génétique supérieure à celle <strong>de</strong>s autres populations et<br />

la différenciation génétique entre populations diminuera d’autant plus que le taux<br />

d’extinction sera élevé (Wa<strong>de</strong> & McCauley 1988, Whitlock & McCauley 1990). Au<br />

contraire, si le nombre <strong>de</strong> colonisateurs est inférieur à 2 fois le nombre <strong>de</strong> migrants entre<br />

populations établies, la diversité génétique <strong>de</strong>s populations recolonisées sera alors plus<br />

faible que dans les autres populations et la différenciation génétique entre populations<br />

augmentera avec le taux d’extinction, comme dans le modèle « propagule ».<br />

Extinctions et recolonisations influencent donc la différenciation génétique entre<br />

populations. Il est dès lors extrêmement difficile <strong>de</strong> relier la différenciation génétique au<br />

<strong>de</strong>gré d’échange entre populations sans connaître au préalable le taux d’extinction appliqué<br />

aux populations. La relation entre Fst et nombre <strong>de</strong> migrants proposée par Wright (1931 ;<br />

Fst = 1/(4Nm+1) doit être adaptée <strong>pour</strong> tenir compte du caractère dynamique <strong>de</strong> la<br />

métapopulation et inclure 2 nouveaux paramètres : le taux d’extinction et la probabilité que<br />

2 colonisateurs aient la même origine (Whitlock & McCauley 1999).<br />

Ajoutées aux capacités <strong>de</strong> dispersion d’une espèce (et donc aux potentialités <strong>de</strong><br />

recolonisation), les caractéristiques <strong>de</strong> l’habitat peuvent également permettre <strong>de</strong> choisir le<br />

modèle <strong>de</strong> dispersion le plus approprié. Ainsi, un habitat extrêmement variable dans le<br />

temps et dans l’espace induira <strong>de</strong>s extinctions et recolonisations <strong>de</strong> populations plus<br />

fréquentes, mieux décrites par un modèle <strong>de</strong> métapopulation.<br />

L’isolement génétique provoqué par une barrière à la dispersion sera donc<br />

différemment ressenti par les espèces en fonction <strong>de</strong> leur modèle <strong>de</strong> population. Selon le<br />

modèle considéré, l’impact d’une barrière à la dispersion <strong>pour</strong>ra ainsi être évalué par <strong>de</strong>s<br />

approches <strong>de</strong> génétique <strong>de</strong>s populations basées sur l’écart à l’équilibre Hardy-Weinberg,<br />

en utilisant les fréquences alléliques et leur distribution dans l’espace sous l’hypothèse<br />

d’un équilibre migration-dérive. Si ces approches en fréquence permettent une vision<br />

relativement récente <strong>de</strong> la différenciation entre populations, la quantification <strong>de</strong> la<br />

8


Introduction générale<br />

divergence accumulée entre ces populations nécessite, quant à elle, une approche différente<br />

basée sur l’accumulation <strong>de</strong>s mutations au cours du temps sous l’hypothèse d’un équilibre<br />

mutation-dérive (échelle <strong>de</strong> temps plus ancienne) et utilisant les arbres phylogénétiques ou<br />

les généalogies d’allèles (ou coalescents).<br />

3. Approche <strong>de</strong> la phylogéographie par la théorie <strong>de</strong> la coalescence<br />

3.1. Principe <strong>de</strong> la théorie <strong>de</strong> la coalescence<br />

À chaque génération, <strong>de</strong>s erreurs <strong>de</strong> réplication <strong>de</strong> l’ADN se produisent. Lorsqu’elles<br />

s’exercent sur les lignées germinales (patrimoine génétique transmis à la <strong>de</strong>scendance), ces<br />

erreurs peuvent être transmises à la génération suivante. Les mutations, définies comme<br />

changement héréditaire dans le matériel génétique, sont <strong>de</strong>s événements rares (10 -4 à 10 -6<br />

nouvelles mutations par gène et par génération) et constituent <strong>pour</strong>tant la source<br />

primordiale <strong>de</strong> variation génétique entre individus. Les évènements <strong>de</strong> mutation se<br />

produisent aléatoirement dans le temps suivant une loi <strong>de</strong> Poisson. Au fil du temps,<br />

certaines <strong>de</strong> ces mutations vont se propager au sein d’une population jusqu’à se fixer avec<br />

un temps moyen <strong>de</strong> 4N (où N est la taille <strong>de</strong> la population diploï<strong>de</strong>) tandis que d’autres<br />

seront éliminées par dérive génétique. À taille <strong>de</strong> population constante, on atteint<br />

rapi<strong>de</strong>ment un état d’équilibre entre le nombre <strong>de</strong> mutations nouvellement apparues et le<br />

nombre <strong>de</strong> mutations qui disparaissent par dérive génétique : c’est l’équilibre mutationdérive<br />

(e.g. Fig. 1.4). De même, un modèle en nombre <strong>de</strong> sites nucléotidiques infini peut,<br />

dans certains cas, être assumé <strong>pour</strong> simplifier les analyses (chaque nouvelle mutation<br />

survient sur un site différent, ce qui a <strong>pour</strong> conséquence <strong>de</strong> ne pas considérer l’existence <strong>de</strong><br />

sites homoplasiques et <strong>de</strong> mutations pouvant retourner à l’état ancestral). Ces modèles<br />

apparaissent néanmoins extrêmement simplificateurs puisqu’il existe <strong>de</strong>s ‘hot spots’ <strong>de</strong><br />

mutations dans les génomes (Li & Graur 1991).<br />

9


Introduction générale<br />

Jukes & Cantor (1969, JC)<br />

Fréquences i<strong>de</strong>ntiques <strong>de</strong>s bases<br />

Probabilités i<strong>de</strong>ntiques d’occurrence <strong>de</strong>s différents types <strong>de</strong> substitution<br />

Biais transition/transversion<br />

Fréquences <strong>de</strong>s bases variables<br />

Kimura 2 paramètres (1980, K2P)<br />

Fréquences i<strong>de</strong>ntiques <strong>de</strong>s bases<br />

Transitions (C↔T et A↔G) plus probables<br />

que les autres substitutions (transversions)<br />

Felsenstein (1981)<br />

Fréquences différentes <strong>de</strong>s bases<br />

Probabilités i<strong>de</strong>ntiques d’occurrence<br />

<strong>de</strong>s différents types <strong>de</strong> substitution<br />

Fréquences <strong>de</strong>s bases variables<br />

Biais transition/transversion<br />

Hasegawa et al (1985, HKY)<br />

Fréquences différentes <strong>de</strong>s bases<br />

Transitions (C↔T et A↔G) plus probables que les transversions (autres combinaisons)<br />

Fréquences <strong>de</strong>s 6 paires <strong>de</strong> substitution pouvant être différentes<br />

Réversible général (Rodríguez et al 1990, REV)<br />

Fréquences différentes <strong>de</strong>s bases<br />

Taux <strong>de</strong> substitution différents <strong>pour</strong> toutes les paires <strong>de</strong> substitution<br />

Fig 1.4 Relation entre cinq modèles <strong>de</strong> substitution. Adaptée <strong>de</strong> Page & Holmes (1998)<br />

10


Introduction générale<br />

Le <strong>de</strong>venir d’une mutation va dépendre <strong>de</strong> l’impact <strong>de</strong>s autres forces évolutives (i.e.<br />

sélection, dérive génétique, migration) sur le site muté, la sélection pouvant également<br />

affecter les éventuelles mutations qui lui sont liées (notion d’autostop moléculaire). Ainsi,<br />

si le temps d’isolement géographique entre <strong>de</strong>ux populations est suffisamment long, <strong>de</strong>s<br />

mutations différentes entre ces populations <strong>pour</strong>ront s’accumuler par sélection<br />

différentielle ou simple dérive génétique, pouvant donner lieu à <strong>de</strong> nouveaux allèles entre<br />

ces populations. Le polymorphisme moléculaire (i.e. basé sur les séquences nucléotidiques<br />

d’ADN) présent s’inscrit donc comme une empreinte <strong>de</strong>s évènements passés, et<br />

notamment <strong>de</strong>s événements d’isolements géographiques. La théorie <strong>de</strong> la coalescence<br />

(Kingman 1982) consiste à utiliser cette empreinte afin <strong>de</strong> « rembobiner le fil <strong>de</strong><br />

l’évolution » et <strong>de</strong> décrypter l’histoire <strong>de</strong>s populations (Fig. 1.4.A). Elle contraste donc<br />

avec les modèles existants jusqu’alors, qui comparent les fréquences alléliques <strong>de</strong>s<br />

populations <strong>de</strong> génération en génération (selon le cours du temps) afin d’en observer les<br />

changements (Wright 1977).<br />

La théorie <strong>de</strong> la coalescence consiste à visualiser les relations entre les lignées actuelles<br />

en remontant dans le temps, jusqu’à trouver l’ancêtre commun le plus récent à l’ensemble<br />

<strong>de</strong> ces lignées (MRCA : Most Recent Common Ancestor, Fig. 1.5). La fusion <strong>de</strong> <strong>de</strong>ux<br />

lignées au niveau d’un MRCA est appelée événement <strong>de</strong> coalescence. Pour une espèce<br />

diploï<strong>de</strong>, la probabilité que <strong>de</strong>ux lignées coalescent à la génération précé<strong>de</strong>nte est <strong>de</strong> 1/2N,<br />

où N est la taille <strong>de</strong> la population. Cette probabilité est la même quelle que soit la<br />

génération considérée. Dans un coalescent neutre (absence <strong>de</strong> sélection et taille <strong>de</strong><br />

population constante), le temps qui sépare <strong>de</strong>ux lignées qui coalescent augmente <strong>de</strong><br />

manière exponentielle lorsque l’on remonte dans le temps, certaines lignées ayant été<br />

perdue au cours <strong>de</strong>s générations (Kingman 1982, T4 < T3 < T2 < T1 sur la Fig. 1.5). Le<br />

coalescent représente ainsi la fusion <strong>de</strong> l’ensemble <strong>de</strong>s lignées échantillonnées. La théorie<br />

<strong>de</strong> la coalescence ne s’intéresse donc qu’aux lignées ayant fournies <strong>de</strong>s <strong>de</strong>scendants actuels<br />

(Fig. 1.5.B). Les lignées ayant disparu ne sont pas prises en considération. La théorie <strong>de</strong> la<br />

coalescence peut ainsi permettre <strong>de</strong> mieux appréhen<strong>de</strong>r les processus ayant conduit à<br />

l’évolution <strong>de</strong>s populations actuelles en étudiant l’écart du coalescent observé par rapport à<br />

un coalescent neutre et la distribution <strong>de</strong>s lignées dans le coalescent au regard <strong>de</strong> la<br />

géographie.<br />

11


Introduction générale<br />

A<br />

Temps en<br />

générations<br />

B<br />

T 1<br />

T 2<br />

T 3<br />

T 4<br />

Fig. 1.5 Processus <strong>de</strong> coalescence sous l’hypothèse neutre.<br />

Chaque cercle représente une lignée. La population est supposée <strong>de</strong> taille constante. En<br />

bleu, lignées actuelles échantillonnées. En rouge, ancêtre commun correspondant à un<br />

événement <strong>de</strong> coalescence. A) Modèle <strong>de</strong> Wright représentant l’ensemble <strong>de</strong>s lignées au<br />

cours du temps. En noir, relation entre lignées ayant données lieu aux lignées actuelles<br />

échantillonnées. En gris, relation entre lignées dont la <strong>de</strong>scendance n’a pas été<br />

échantillonnée. B) Modèle <strong>de</strong> coalescence <strong>de</strong> Kingman ne prenant en compte que les lignées<br />

ayant engendrées une <strong>de</strong>scendance échantillonnée.<br />

12


Introduction générale<br />

3.2. Écart à la neutralité : impact <strong>de</strong> l’histoire démographique<br />

et <strong>de</strong> la sélection naturelle sur le coalescent<br />

L’étu<strong>de</strong> <strong>de</strong> la distribution <strong>de</strong>s mutations sur un coalescent peut apporter <strong>de</strong>s<br />

informations précieuses sur les effets démographiques et/ou sélectifs pouvant intervenir<br />

dans les populations. En absence <strong>de</strong> sélection et d’événement démographique (taille <strong>de</strong><br />

population constante), les mutations s’accumulent aléatoirement dans le polymorphisme et<br />

la divergence d’un gène. La longueur <strong>de</strong>s branches sur un coalescent étant proportionnelle<br />

au temps, les branches internes auront donc <strong>de</strong>s longueurs supérieures aux branches<br />

terminales (Fig. 1.6.A) puisque ayant accumulé plus <strong>de</strong> mutations (égal au nombre <strong>de</strong><br />

générations nécessaires <strong>pour</strong> arriver au MRCA).<br />

Supposons maintenant qu’un événement démographique ait engendré une<br />

augmentation <strong>de</strong> la taille <strong>de</strong> la population à un temps t. La population observée<br />

actuellement (No) présente alors une taille supérieure à la taille <strong>de</strong> la population ancestrale<br />

au temps t (Nt). Deux lignées ancestrales (plus anciennes que t) présentent alors une<br />

probabilité plus gran<strong>de</strong> <strong>de</strong> coalescer à la génération précé<strong>de</strong>nte que <strong>de</strong>ux lignées actuelles<br />

(1/2No < 1/2Nt). En conséquence, le temps nécessaire <strong>pour</strong> que <strong>de</strong>ux lignées actuelles<br />

coalescent sera plus long qu’entre lignées ancestrales et donc les branches terminales du<br />

coalescent ayant subi une expansion démographique seront plus longues que les branches<br />

internes (Fig. 1.6.B). Le coalescent aura une forme <strong>de</strong> râteau ou d’étoile (Hudson &<br />

Slatkin 1991). Les mutations auront donc plus <strong>de</strong> chance <strong>de</strong> s’accumuler sur les branches<br />

terminales. On observera un excès <strong>de</strong> variants rares (mutations en faible fréquence ou<br />

n’étant échantillonnées qu’une seule fois : singletons). À l’inverse, une diminution <strong>de</strong> taille<br />

<strong>de</strong> population engendrera <strong>de</strong>s branches internes plus courtes que les branches terminales,<br />

avec excès <strong>de</strong> variants en fréquence intermédiaire. Une expansion démographique faisant<br />

généralement suite à une réduction <strong>de</strong> population (effet fondateur), la topologie observée<br />

sera alors un râteau avec une branche interne relativement longue (Fig. 1.6.C).<br />

13


Introduction générale<br />

A<br />

B<br />

C<br />

Fig. 1.6 Coalescents simulés selon (A) l’hypothèse neutre, (B) un facteur <strong>de</strong> croissance<br />

démographique <strong>de</strong> 1000 à l’ai<strong>de</strong> du logiciel TreeToy 1.b, (C) coalescent attendu lors<br />

d’un goulôt d’étranglement suivi d’une expansion démographique.<br />

Outre les effets démographiques, les gènes sous sélection vont également présenter <strong>de</strong>s<br />

topologies <strong>de</strong> coalescent qui s’écartent du coalescent neutre. Les différents types <strong>de</strong><br />

sélection peuvent ainsi se classer en <strong>de</strong>ux catégories selon qu’elles engendrent une<br />

réduction ou un maintien du polymorphisme.<br />

La sélection positive consiste en la fixation d’un allèle avantageux dans la population<br />

diminuant ainsi la diversité génétique <strong>de</strong> manière drastique lors <strong>de</strong> l’augmentation en<br />

fréquence <strong>de</strong> la mutation avantageuse dans les générations qui suivent son apparition. De<br />

nouvelles mutations <strong>pour</strong>ront alors s’accumuler dans le polymorphisme lorsque l’allèle<br />

avantageux aura remplacé toutes les autres lignées ancestrales (notion <strong>de</strong> balayage<br />

sélectif). Le coalescent présentera une topologie en râteau similaire à celle obtenue lors<br />

d’un goulot d’étranglement démographique (réduction drastique <strong>de</strong> la taille <strong>de</strong> la<br />

population et donc <strong>de</strong> la diversité puis expansion, Fig. 1.6.C). Certains gènes peuvent ne<br />

pas être directement sous sélection positive (mutation adaptative apparue ailleurs dans un<br />

gène proche) mais être entraînés dans le balayage (par effet autostop ou hitch-hiking) avec<br />

le gène qui possè<strong>de</strong> l’allèle avantageux. Ces gènes seront donc en déséquilibre <strong>de</strong> liaison<br />

avec le gène porteur <strong>de</strong> l’avantage adaptatif.<br />

14


Introduction générale<br />

La sélection purificatrice (négative) conduit également à une réduction <strong>de</strong> diversité<br />

génétique en éliminant les mutations délétères (ou les mutations neutres ou faiblement<br />

avantageuses qui lui sont liées par l’effet d’un autostop : sélection d’arrière-plan ou<br />

background selection). Par définition, elle affecte tous les gènes ayant une fonction dans<br />

l’organisme et <strong>pour</strong> lesquels la structure <strong>de</strong> la molécule transcrite et/ou traduite nécessite<br />

d’être conservée. De manière générale sur les parties codantes d’un gène, elle aura <strong>pour</strong><br />

conséquence <strong>de</strong> réduire par trois la diversité génétique attendue sous l’évolution neutre<br />

<strong>pour</strong> une taille <strong>de</strong> population donnée (voir Li & Graur 1991). Néanmoins, son action dans<br />

le temps est plus diffuse (pas <strong>de</strong> diminution rapi<strong>de</strong> <strong>de</strong> diversité génétique) que celle<br />

attendue lors d’une propagation d’un allèle avantageux. En conséquence, il est<br />

extrêmement difficile <strong>de</strong> distinguer l’action <strong>de</strong> la sélection purificatrice (ou sélection<br />

d’arrière-plan) d’un taux <strong>de</strong> mutation faible <strong>pour</strong> un gène donné ou d’un effet positif (ou<br />

balayage sélectif) extrêmement faible <strong>de</strong> la sélection (Charlesworth et al 1993). Le<br />

coalescent aura une topologie très semblable à une topologie neutre mais avec un nombre<br />

beaucoup plus réduit d’allèles.<br />

À l’inverse, différents allèles d’un même gène soumis à la sélection balancée tendront<br />

à être conservés pendant une pério<strong>de</strong> <strong>de</strong> temps plus longue que <strong>de</strong>s allèles neutres. La<br />

sélection balancée engendrera donc un maintien <strong>de</strong> polymorphisme plus important que<br />

sous la neutralité et présentera ainsi un coalescent avec <strong>de</strong>s branches plus longues que dans<br />

un coalescent neutre (Fig. 1.7).<br />

15


Introduction générale<br />

Fig. 1.7 Topologies d’arbres <strong>de</strong> coalescence (A) d’un gène neutre, (B) d’un gène<br />

soumis à sélection balancée.<br />

En pointillés, racine <strong>de</strong> l’arbre.<br />

Si la topologie d’un coalescent seule ne permet pas, à elle seule, d’affirmer un écart<br />

à une accumulation neutre du polymorphisme, elle constitue un indice précieux permettant<br />

<strong>de</strong> poser <strong>de</strong>s hypothèses pouvant être testées par le biais <strong>de</strong> tests statistiques. Un effet<br />

démographique ou sélectif pouvant avoir le même impact sur la topologie du coalescent<br />

d’un gène, il peut être difficile <strong>de</strong> discriminer ces causes. Pour cela, il est alors nécessaire<br />

<strong>de</strong> procé<strong>de</strong>r par une approche comparative : (1) un gène et plusieurs espèces si l’on<br />

soupçonne que l’environnement ait pu affecter la dynamique <strong>de</strong>s populations d’une façon<br />

semblable au sein d’un peuplement (e.g. effet vicariant) ou (2) une espèce et plusieurs<br />

gènes. En effet, un événement démographique affecte l’ensemble <strong>de</strong>s gènes d’une espèce<br />

<strong>de</strong> la même façon, ce qui n’est pas le cas <strong>de</strong> la sélection. Au contraire, la sélection<br />

n’affectera qu’un nombre limité <strong>de</strong> gènes (par convergence adaptative) au sein d’une<br />

espèce, ou d’un petit nombre d’espèces.<br />

3.3. Structure du coalescent au regard <strong>de</strong> la géographie<br />

Outre les effets sélectifs et démographiques, la topologie du coalescent peut<br />

également fournir <strong>de</strong> nombreuses informations sur les processus à l’origine <strong>de</strong> la<br />

distribution géographique actuelle <strong>de</strong> la lignée. Faire figurer sur le coalescent la<br />

localisation géographique <strong>de</strong>s lignées échantillonnées permet <strong>de</strong> visualiser la<br />

correspondance possible entre évènements <strong>de</strong> divergence et région d’échantillonnage,<br />

pouvant révéler la présence <strong>de</strong> barrière aux flux <strong>de</strong> gènes liée à un isolement géographique<br />

(barrière à la dispersion, distance entre sites, voire habitats ou micro-habitats contrastés).<br />

16


Introduction générale<br />

Nombre d’exemples existent dans la littérature ayant mis en évi<strong>de</strong>nce <strong>de</strong>s barrières<br />

géographiques (voir par exemple les revues <strong>de</strong> Maggs et al 2008, Nord Est <strong>de</strong><br />

l’Atlantique ; Soltis et al 2006 Nord Est <strong>de</strong> l’Amérique). La force <strong>de</strong> l’isolement génétique<br />

dépendra <strong>de</strong> la durée et <strong>de</strong> la force <strong>de</strong> l’isolement géographique.<br />

Une barrière à la dispersion (relativement récente ou faible) <strong>pour</strong>ra ainsi engendrer<br />

<strong>de</strong>s différences <strong>de</strong> fréquences génétiques <strong>de</strong> part et d’autre <strong>de</strong> cette barrière. Une approche<br />

couplant coalescence et fréquences alléliques constitue un moyen élégant <strong>de</strong> détecter ce<br />

type <strong>de</strong> barrière phylogéographique. Outre la représentation <strong>de</strong> la coalescence utilisant <strong>de</strong>s<br />

arbres phylogénétiques, la représentation via <strong>de</strong>s réseaux <strong>de</strong> séquences permet <strong>de</strong> visualiser<br />

plus facilement les fréquences <strong>de</strong>s différents allèles, chaque cercle étant proportionnel au<br />

nombre <strong>de</strong> séquences i<strong>de</strong>ntiques échantillonnées. Par exemple, Goldstien et al (2006)<br />

mettent en évi<strong>de</strong>nce la présence d’une barrière phylogéographique chez trois espèces <strong>de</strong><br />

patelles du genre Cellana qui correspond à l’actuelle séparation <strong>de</strong>s îles nord et sud <strong>de</strong> la<br />

Nouvelle Zélan<strong>de</strong> (Fig 1.8). Ces auteurs attribuent l’isolement géographique au complexe<br />

hydrographique dans la région <strong>de</strong> Cook Strait. En couplant réseaux d’haplotypes et<br />

distribution géographique <strong>de</strong>s haplotypes sur le gène mitochondrial <strong>de</strong> la Cytochrome b,<br />

les auteurs ont ainsi montré que les haplotypes les plus fréquents, <strong>pour</strong>tant différents entre<br />

eux d’une mutation, présentaient <strong>de</strong>s fréquences différentes entre les <strong>de</strong>ux îles, avec <strong>de</strong>s<br />

valeurs <strong>de</strong> différenciation génétique variable selon l’espèce (Φst <strong>de</strong> 0,829, 0,142 et 0,177<br />

<strong>pour</strong> C. ornata, C. radians et C. flava respectivement).<br />

Une barrière à la dispersion plus intense ou plus ancienne peut aussi engendrer une<br />

divergence sensiblement équivalente entre lignées phylogéographiques au sein d’une<br />

même espèce (e.g. Fig. 1.9, <strong>de</strong> Moritz & Faith 1998), voire <strong>de</strong> la co-spéciation (e.g. Dooh<br />

et al 2006, Jolly et al 2006).<br />

17


Introduction générale<br />

A<br />

B<br />

Fig. 1.8 Illustration du couplage entre approches issues <strong>de</strong> la théorie <strong>de</strong> la coalescence<br />

(réseaux d’haplotypes) et <strong>de</strong> la génétique <strong>de</strong>s populations « classique » (fréquences<br />

haplotypiques) sur 3 espèces du genre Cellana. Tirée <strong>de</strong> Goldstien et al (2006).<br />

(A) Carte géographique soulignant les caractéristiques hydrographiques <strong>de</strong> la Nouvelle<br />

Zélan<strong>de</strong>. (B) réseaux d’haplotypes et distribution <strong>de</strong> ces haplotypes, les couleurs <strong>de</strong>s<br />

haplotypes sur les réseaux correspon<strong>de</strong>nt à celles présentent sur les cartes.<br />

18


Introduction générale<br />

Gnypetoscincus<br />

queenslandiae<br />

Orthonyx spaldingii<br />

Poecilodryas albispecularis<br />

Fig. 1.9 Détection d’une barrière aux flux <strong>de</strong> gènes via la correspondance entre<br />

géographie et divergence sur les coalescents. Tirée <strong>de</strong> Moritz & Faith (1998)<br />

Arbres phylogénétiques réalisés chez le reptile Gnypetoscincus queenslandiae et les<br />

oiseaux Orthonyx spaldingii et Poecilodryas albispecularis au nord est <strong>de</strong> la forêt<br />

tropicale Queensland (Australie). WT, Windsor Tableland ; CT, Carbine Tableland ; AT,<br />

Atherton Tableland ; CR Cardwell Range. Les nombres correspon<strong>de</strong>nt à l’estimation <strong>de</strong> la<br />

distance <strong>de</strong> divergence <strong>de</strong> Kimura-2-P x1000.<br />

Si le temps <strong>de</strong> divergence séparant <strong>de</strong>ux lignées peut permettre d’avoir une<br />

estimation <strong>de</strong> la date <strong>de</strong> l’isolement géographique, la persistance d’évènements <strong>de</strong><br />

migration, même restreints, après l’initiation d’une barrière géographique, peut engendrer<br />

une mauvaise estimation du début <strong>de</strong> l’isolement (Nielsen & Wakeley 2001). Beaucoup <strong>de</strong><br />

modèles considèrent que les populations géographiquement distinctes : (1) échangent <strong>de</strong>s<br />

migrants à un taux constant pendant un temps infini ou (2) ont divergé à partir d’une<br />

population ancestrale à un temps t, avec absence totale <strong>de</strong> flux <strong>de</strong> gènes entre ces<br />

19


Introduction générale<br />

populations géographiques <strong>de</strong>puis t. Il est clair que le processus d’isolement est souvent<br />

progressif avec <strong>de</strong>s possibilités <strong>de</strong> retour en arrière (notion <strong>de</strong> zones <strong>de</strong> contact<br />

secondaire). De fait, la perméabilité <strong>de</strong> la barrière aux flux <strong>de</strong> gènes peut être incomplète<br />

Une divergence observée faible mais robuste sur un coalescent peut ainsi être due soit à (1)<br />

une barrière à la dispersion complète (pas <strong>de</strong> migration <strong>de</strong>puis l’isolement) mais récente ou<br />

(2) à une barrière à la dispersion plus ancienne mais plus perméable entraînant <strong>de</strong> la<br />

migration sporadique entre les populations isolées et la création d’allèles recombinés entre<br />

formes parentales (Nielsen & Wakeley 2001). Si l’isolement génétique <strong>de</strong>s populations<br />

s’est effectué en présence <strong>de</strong> migration, le temps <strong>de</strong> divergence estimé en considérant<br />

l’absence <strong>de</strong> migration ne correspond alors pas au début du phénomène d’isolement<br />

géographique (temps sous estimé). Des modèles d’isolement avec migration ont donc été<br />

récemment introduits <strong>pour</strong> tenir compte <strong>de</strong>s paramètres <strong>de</strong> migration dans l’estimation <strong>de</strong>s<br />

temps <strong>de</strong> divergence (Nielsen & Wakeley 2001). La position géographique <strong>de</strong> la barrière et<br />

sa largeur peuvent également varier en fonction <strong>de</strong>s capacités <strong>de</strong> dispersion <strong>de</strong>s espèces,<br />

définissant <strong>de</strong> véritables zones <strong>de</strong> transition entre ces régions phylogéographiques.<br />

3.4. Coalescence <strong>de</strong> gènes versus séparation <strong>de</strong>s lignées intraspécifiques<br />

Si la coalescence est un outil puissant permettant notamment d’appréhen<strong>de</strong>r plus<br />

aisément l’évolution d’un gène, la topologie d’un coalescent basée sur un gène ne<br />

correspond pas nécessairement à celle <strong>de</strong> l’espèce (ou <strong>de</strong>s lignées intraspécifiques, Hudson<br />

1983, Pamilo & Nei 1988, Maddison 1997). Ainsi le coalescent d’un gène A peut<br />

correspondre à la topologie réelle entre lignées intraspécifiques alors que celui d’un gène B<br />

présentera une topologie différente (Fig. 1.10). Cela peut être dû à <strong>de</strong>s phénomènes <strong>de</strong><br />

duplication du gène dans une <strong>de</strong>s lignées, la perte du gène dans une <strong>de</strong>s lignées mais<br />

également à la stochasticité du processus <strong>de</strong> coalescence, notamment si celui-ci est<br />

exacerbé par <strong>de</strong>s changements démographiques locaux. Lorsqu’une seule séquence est<br />

échantillonnée par lignée, la probabilité qu’un gène ne présente pas une topologie<br />

concordante avec la topologie réelle a été évaluée, par Pamilo & Nei (1988), à (2/3) e -T , où<br />

T correspond au temps séparant l’ancêtre commun le plus récent (MRCA) entre les lignées<br />

les plus proches et le MRCA à toutes les lignées (Fig. 1.10).<br />

20


Introduction générale<br />

T<br />

Fig. 1.10 Coalescence <strong>de</strong> gène versus coalescence <strong>de</strong> lignées intraspécifiques.<br />

En bleu, représentation <strong>de</strong> la coalescence réelle <strong>de</strong>s lignées intraspécifiques. En rouge,<br />

coalescent d’un gène A en accord avec la topologie réelle <strong>de</strong> ces lignées. En vert,<br />

coalescent d’un gène B ne reflétant pas la coalescence réelle <strong>de</strong> ces lignées.<br />

Pour illustrer cette différence entre coalescence <strong>de</strong> gènes et coalescence d’espèces,<br />

prenons l’exemple <strong>de</strong>s mandarins diamants du genre Poephila. Jennings & Edwards (2005)<br />

ont effectué <strong>de</strong>s coalescents <strong>de</strong> 30 gènes indépendants sur ces oiseaux et ont mis en<br />

évi<strong>de</strong>nce une divergence génétique entre espèces <strong>pour</strong> 28 d’entre eux. Si 16 <strong>de</strong> ces locus<br />

montraient un regroupement <strong>de</strong> P. acuticauda (P.a.) et P. hecki (P.h.) par rapport à P.<br />

cincta (P.c.), soit une topologie : ((P.a., P.h.), P.c.) considérée comme représentative <strong>de</strong> la<br />

réalité ; 7 autres locus présentaient une topologie ((P.a., P.c.), P.h.) et les 5 locus restant<br />

une topologie ((P.c., P.h.), P.a.). Le calcul <strong>de</strong> la divergence T à partir <strong>de</strong>s données<br />

observées, ne considérant que les 16 premiers locus, est concordant avec le T simulé sur<br />

l’ensemble <strong>de</strong>s locus en considérant que ces 16 locus concor<strong>de</strong>nt avec la topologie réelle et<br />

que les autres locus constituent un ‘pool’ <strong>de</strong> gènes s’écartant par dérive du coalescent<br />

‘vrai’ (Wakeley 2009). Il est donc nécessaire <strong>de</strong> comparer les topologies <strong>de</strong> plusieurs gènes<br />

et <strong>de</strong> rester critique quant aux coalescents obtenus sur un faible nombre <strong>de</strong> gènes <strong>pour</strong><br />

expliquer l’histoire évolutive d’une espèce.<br />

21


Introduction générale<br />

4. Relâchement d’un isolement géographique<br />

Les barrières à la dispersion se mettent en place progressivement (e.g. fermeture <strong>de</strong><br />

l’Isthme <strong>de</strong> Panama : Knowlton 2000) et ne sont pas immuables dans le temps. Ainsi, si la<br />

mise en place d’une barrière peut engendrer un isolement géographique puis génétique <strong>de</strong>s<br />

populations, le relâchement <strong>de</strong> sa perméabilité peut permettre <strong>de</strong> nouveaux échanges entre<br />

lignées dès lors que l’isolement reproductif n’est pas atteint, et favoriser <strong>de</strong>s zones <strong>de</strong><br />

contact secondaire entre les entités taxonomiques précé<strong>de</strong>mment séparées (e.g. contact<br />

secondaire au niveau du Cap Canaveral chez le crabe du genre Menippe : Liu et al 1991,<br />

ou le gastéropo<strong>de</strong> du genre Stramonita : Bert & Arnold 1995). Si cet isolement <strong>de</strong>s<br />

populations a été suffisamment long, ces lignées peuvent ne plus être interfécon<strong>de</strong>s et<br />

seront alors considérées comme <strong>de</strong>s espèces à part entière. Il en résulte alors une simple<br />

juxtaposition <strong>de</strong>s lignées dans les zones <strong>de</strong> sympatrie. Dans le cas contraire, la remise en<br />

contact secondaire peut être suivie par la formation d’individus hybri<strong>de</strong>s entre lignées<br />

génétiques, voire d’une introgression d’allèles d’un génome dans l’autre, si l’individu<br />

hybri<strong>de</strong> est capable <strong>de</strong> se croiser avec les formes parentales. S’ils sont défavorisés, une<br />

sélection contre les hybri<strong>de</strong>s va tendre à augmenter l’isolement reproductif, s’opposant<br />

ainsi à la migration (force homogénéisante). L’équilibre entre isolement reproductif et<br />

migration formera alors une zone d’hybridation (Barton 1979, Barton & Hewitt 1985),<br />

avec la mise en place d’un gradient <strong>de</strong> fréquences alléliques (cline) autour <strong>de</strong> la barrière<br />

géographique ayant engendré la divergence entre les lignées. L’intensité du cline sera<br />

variable en fonction <strong>de</strong> l’implication du gène étudié dans le processus d’isolement<br />

reproductif (ou du <strong>de</strong>gré <strong>de</strong> liaison <strong>de</strong>s gènes avec ces gènes « verrous », Kruuk et al 1999,<br />

Fig. 1.11).<br />

22


Introduction générale<br />

Fréquence allélique<br />

0<br />

Distance par rapport à la barrière aux flux <strong>de</strong> gènes<br />

Fig. 1.11 Intensité du cline géographique en fonction <strong>de</strong> l’implication du gène dans<br />

l’isolement reproductif <strong>de</strong>s espèces situées <strong>de</strong> part et d’autre <strong>de</strong> la barrière. Adaptée <strong>de</strong><br />

Kruuk et al (1999)<br />

En pointillés, figuration <strong>de</strong> la barrière aux flux <strong>de</strong> gènes. Les courbes représentent le cline<br />

<strong>de</strong> fréquences alléliques <strong>pour</strong> trois gènes. Plus l’épaisseur <strong>de</strong> la courbe est importante, plus<br />

le gène est impliqué dans l’isolement reproductif (ou lié à un gène impliqué).<br />

23


Introduction générale<br />

5. Biogéographie en milieu hydrothermal profond<br />

Au fond <strong>de</strong>s océans, les mouvements tectoniques <strong>de</strong>s plaques ont conduit à la<br />

formation <strong>de</strong> dorsales océaniques, lieu <strong>de</strong> l’expansion <strong>de</strong> la croûte océanique par remontée<br />

<strong>de</strong> magma basaltique. À ce niveau, l’eau <strong>de</strong> mer froi<strong>de</strong> s’infiltre par les fissures associées<br />

aux forces <strong>de</strong> cisaillement et pénètre jusqu’à la chambre magmatique. Elle se charge alors<br />

en métaux lourds, en gaz et en composés réduits, s’échauffe et remonte en surface<br />

(Zierenberg et al 2000). Cette eau transformée reste en phase liqui<strong>de</strong> grâce aux pressions<br />

hydrauliques importantes présentes à ces profon<strong>de</strong>urs. Elle constitue le flui<strong>de</strong><br />

hydrothermal. Le contact entre ce flui<strong>de</strong> et l’eau froi<strong>de</strong> du fond <strong>de</strong>s océans précipite les<br />

éléments métalliques et forme alors les cheminées hydrothermales. Géologues et<br />

géochimistes ont longtemps débattu sur la présence <strong>de</strong> flui<strong>de</strong> hydrothermal potentiel<br />

associé aux dorsales océaniques (zones d’accrétion), et ont ainsi effectué <strong>de</strong>s campagnes<br />

océanographiques explorant ces dorsales à la recherche d’infiltration d’eau profon<strong>de</strong> dans<br />

le plancher océanique. Si la vie en absence <strong>de</strong> lumière et avec <strong>de</strong>s pressions hydrostatiques<br />

importantes fût mise en évi<strong>de</strong>nce dans les années 50 (faune abyssale : Fukushima 2000),<br />

les propriétés physico-chimiques <strong>de</strong> ce flui<strong>de</strong> (fortes températures, pH aci<strong>de</strong>, anoxie, fortes<br />

concentrations en H 2 S) laissaient penser à l’impossibilité d’une forme <strong>de</strong> vie associée au<br />

flui<strong>de</strong> hydrothermal. Pourtant, c’est bien là, associée à ces cheminées actives, que le 15<br />

février 1977, <strong>de</strong>s géochimistes et géologues américains ont découvert la présence d’une<br />

faune foisonnante <strong>de</strong> vers géants et <strong>de</strong> bivalves sur la ri<strong>de</strong> <strong>de</strong>s Galápagos (Lonsdale 1977).<br />

Bien que les organismes hydrothermaux consomment <strong>de</strong> l’O 2 (en provenance <strong>de</strong> la<br />

surface <strong>de</strong>s océans) <strong>pour</strong> vivre, cet écosystème ne dépend pas directement <strong>de</strong> la lumière et<br />

<strong>de</strong> la production primaire photosynthétique. Il est essentiellement basé sur la<br />

chimiosynthèse effectuée par <strong>de</strong>s bactéries vivant libres ou associées aux métazoaires sous<br />

forme <strong>de</strong> symbiose (Childress & Fisher 1992). C’est <strong>pour</strong>quoi, les espèces hydrothermales<br />

se répartissent <strong>de</strong> manière concentrique autour <strong>de</strong>s émissions <strong>de</strong> flui<strong>de</strong> hydrothermal,<br />

source <strong>de</strong>s éléments réduits nécessaires à cette chimiosynthèse bactérienne. Cette<br />

organisation spatiale particulière dépend <strong>de</strong>s besoins trophiques <strong>de</strong>s animaux, <strong>de</strong> leurs<br />

capacités d’adaptation à la toxicité du flui<strong>de</strong> et <strong>de</strong>s compétitions inter-spécifiques <strong>pour</strong> la<br />

ressource. Ainsi, les communautés hydrothermales forment <strong>de</strong> véritables « oasis » (Laubier<br />

24


Introduction générale<br />

1986) séparés parfois par plusieurs dizaines <strong>de</strong> kilomètres <strong>de</strong> « désert » abyssal. Elles sont<br />

donc entièrement dépendantes <strong>de</strong>s sources hydrothermales actives et, <strong>de</strong> ce fait, la plupart<br />

<strong>de</strong>s espèces formant ces communautés sont endémiques <strong>de</strong> ce milieu et disparaissent à<br />

l’arrêt <strong>de</strong>s émissions.<br />

Or, la durée <strong>de</strong> vie <strong>de</strong> ces sources hydrothermales est limitée dans le temps et<br />

dépend <strong>de</strong> la vitesse d’accrétion <strong>de</strong> la dorsale sur laquelle elles se forment. Les fissures par<br />

lesquelles l’eau ressort sous la forme du flui<strong>de</strong> hydrothermal peuvent s’obstruer, par<br />

colmatage métallique <strong>de</strong>s conduits provoquant ainsi le tarissement <strong>de</strong> la source. D’autres<br />

fissures se reformeront plus ou moins loin <strong>de</strong> la source éteinte, donnant ainsi naissance à<br />

une nouvelle source active. L’instabilité du milieu est donc à la fois spatiale et temporelle.<br />

Plus l’accrétion est rapi<strong>de</strong>, plus le milieu est instable. Les espèces connaissent ainsi <strong>de</strong><br />

véritables phases d’extinction par <strong>de</strong>struction rapi<strong>de</strong> <strong>de</strong> l’habitat (épanchements <strong>de</strong> lave sur<br />

les sites d’activité) affectant souvent la quasi-totalité d’un champ hydrothermal (Haymon<br />

et al 1991, Jollivet 1993, Tunnicliffe et al 1997). Ces phases d’extinction sont très<br />

rapi<strong>de</strong>ment suivies d’un ‘bloom’ bactérien et d’une recolonisation massive <strong>de</strong> l’habitat,<br />

d’abord par la faune vagile (crabes, galathées) puis par la faune sessile (vestimentifères)<br />

avec <strong>de</strong>s taux <strong>de</strong> croissance extrêmement élevés (Lutz et al 1994, Shank et al 1998) faisant<br />

<strong>de</strong> ce milieu, un cas exemplaire <strong>de</strong> fonctionnement en métapopulation (Vrijenhoek 1997,<br />

Jollivet et al 1999).<br />

Les sources hydrothermales sont présentes dans tous les océans (Fig. 1.12). Si <strong>de</strong><br />

nombreux sites hydrothermaux ont d’ores et déjà été découverts, le système <strong>de</strong> dorsale<br />

médio-océanique représente près <strong>de</strong> 60 000 km <strong>de</strong> long et reste encore largement<br />

inexploré. Pratiquement aucune mission n’a été effectuée le long <strong>de</strong>s dorsales Antarctique<br />

et Indienne (Fig. 1.12), bien que la présence <strong>de</strong> sources hydrothermales y ait été détectée<br />

récemment (Bohrman et al 1999). Il n’est ainsi pas rare <strong>de</strong> mettre à jour <strong>de</strong> nouvelles<br />

sources actives lors <strong>de</strong> plongées sur <strong>de</strong>s zones encore non visitées, la <strong>de</strong>rnière découverte<br />

en date a ainsi été effectuée en août/septembre 2008 sur la dorsale du Pacifique oriental (ou<br />

East Pacific Rise, EPR ; 102,6º Ouest / 2,1º Sud) lors <strong>de</strong> la mission chinoise DY115-20.<br />

25


Introduction générale<br />

Fig. 1.12 Localisation <strong>de</strong>s sources hydrothermales et <strong>de</strong>s provinces biogéographiques. Adaptée <strong>de</strong> Desbruyères et al (2006b)<br />

En cercle noir, les provinces biogéographiques généralement admises : 1, dorsale pacifique orientale ; 2, dorsale pacifique nord ; 3, dorsale<br />

médio-atlantique ; 4, bassins arrière arc du pacifique occi<strong>de</strong>ntal ; 5, sources hydrothermales <strong>de</strong> l’océan Indien.<br />

En cercle bleu, les provinces biogéographiques proposées par certains auteurs (voir texte) : a, dorsale pacifique orientale nord et Galápagos ;<br />

b, dorsale pacifique orientale sud ; c, région <strong>de</strong>s Azores ; d, dorsale médio-atlantique au sud <strong>de</strong>s Azores ; e, pacifique occi<strong>de</strong>ntal nord ; f,<br />

pacifique occi<strong>de</strong>ntal sud et sources hydrothermales <strong>de</strong> l’océan Indien<br />

26


Introduction générale<br />

Depuis la découverte <strong>de</strong> la faune hydrothermale, le nombre d’espèces décrites n’a<br />

cessé d’augmenter (533 recensées dans Desbruyères et al 2006b, mais <strong>de</strong>s dizaines d’autres<br />

espèces sont actuellement en cours d’i<strong>de</strong>ntification). Leur distribution géographique a<br />

rapi<strong>de</strong>ment suscité l’intérêt <strong>de</strong>s biogéographes. Malgré la très gran<strong>de</strong> homogénéité <strong>de</strong>s<br />

conditions hydrothermales à l’échelle du globe, la composition faunistique du milieu<br />

hydrothermal présente en effet <strong>de</strong> nombreuses disparités géographiques. L’ensemble <strong>de</strong>s<br />

auteurs s’accor<strong>de</strong> <strong>pour</strong> définir au moins cinq provinces biogéographiques différentes<br />

(Tunnicliffe 1991, Tunnicliffe et al 1998, Van Dover et al 2002, Tyler et al 2003, Fig.<br />

1.12).<br />

L’EPR fut la première province biogéographique i<strong>de</strong>ntifiée. Sa composition<br />

faunistique (Fig. 1.13) est dominée en biomasse par les vers géants Riftia pachyptila, les<br />

bivalves Bathymodiolus thermophilus et Calyptogena magnifica et les espèces <strong>de</strong><br />

polychètes Alvinellidae (telles que le vers <strong>de</strong> pompéi Alvinella pompejana, animal résistant<br />

à <strong>de</strong>s températures élevées, pouvant atteindre, pendant <strong>de</strong> courtes durées, 105°C,<br />

Chevaldonné et al 1992) et, en nombre, par <strong>de</strong>s gastéropo<strong>de</strong>s <strong>de</strong> petite taille, notamment<br />

les Lepetodrilidae et les Peltospiridae (Hessler et Smithey 1983, Fustec et al 1987).<br />

La secon<strong>de</strong> province biogéographique décrite fut la dorsale du Nord-Est Pacifique.<br />

Cette province est composée principalement par une faune similaire à celle <strong>de</strong> l’EPR mais<br />

composée <strong>de</strong> genres et/ou <strong>de</strong> familles proches, tels que le ver géant Ridgeia piscesae, les<br />

polychète Alvinellidae Paralvinella sulfincola et P. palmiformis (Tunnicliffe et al 1985,<br />

1986, Fig. 1.13). Il s’agit <strong>de</strong> la seule province biogéographique décrite qui ne compte pas<br />

<strong>de</strong> moules hydrothermales du genre Bathymodiolus (Tyler et al 2003, les moules décrites<br />

sur cette province sont <strong>de</strong>s genres Benthomodiolus à Gorda et Idas et Adipicola sur Juan <strong>de</strong><br />

Fuca). La subduction <strong>de</strong> la plaque Pacifique sous la plaque Farallon il y a environ 35<br />

millions d’années, aurait divisé la dorsale océanique <strong>de</strong> part et d’autre <strong>de</strong> l’Amérique du<br />

Nord, engendrant un événement <strong>de</strong> vicariance majeur ayant conduit à la différenciation <strong>de</strong><br />

ces <strong>de</strong>ux provinces biogéographiques par la mise en place <strong>de</strong> faunes jumelles (Tunnicliffe<br />

1988).<br />

La dorsale médio-Atlantique constitue la troisième province biogéographique mise<br />

en évi<strong>de</strong>nce, probablement séparée <strong>de</strong>s provinces pacifiques lors <strong>de</strong> la fermeture <strong>de</strong><br />

l’isthme <strong>de</strong> Panama (∼10 Ma, Van Dover et al 2002). En plus <strong>de</strong>s moulières <strong>de</strong><br />

Bathymodiolus azoricus et B. puteoserpentis, la composition faunistique <strong>de</strong> cette province<br />

contraste fortement avec les <strong>de</strong>ux précé<strong>de</strong>ntes. Ainsi, les vestimentifères tubicoles sont<br />

27


Introduction générale<br />

absents et <strong>de</strong> véritables essaims <strong>de</strong> crevettes <strong>de</strong> l’espèce Rimicaris exoculata recouvrent<br />

parfois certaines cheminées hydrothermales (Desbruyères et al 2000, Fig. 1.13). On<br />

retrouve cependant une dominance numérique <strong>de</strong> certains petits gastéropo<strong>de</strong>s (Lepetodrilus<br />

atlanticus, Protolyra valvatoi<strong>de</strong>s) à certains endroits <strong>de</strong> la dorsale.<br />

La faune hydrothermale <strong>de</strong>s bassins arrière arc du pacifique occi<strong>de</strong>ntal est<br />

également dominée par les bivalves du genre Bathymodiolus (i.e. B. brevior, B.<br />

septemdierum, B. elongatus) mais également par <strong>de</strong>s gastéropo<strong>de</strong>s symbiotiques <strong>de</strong> gran<strong>de</strong><br />

taille appartenant à la famille <strong>de</strong>s Provannidae tels que Alviniconcha hessleri et Ifremeria<br />

nautilei (Desbruyères et al 1994, Fig. 1.13).<br />

Enfin, découverts récemment, les sites hydrothermaux <strong>de</strong> l’océan Indien semblerait<br />

constituer une cinquième province biogéographique, bien que <strong>de</strong> nombreuses espèces<br />

restent encore à décrire et que les faunes paraissent intermédiaires entre la zone Atlantique<br />

et celle du Pacifique occi<strong>de</strong>ntal (Van Dover et al 2001a). Cette province est dominée par<br />

<strong>de</strong>s crevettes Bresiliidae (proche <strong>de</strong> celles trouvées en Atlantique), <strong>de</strong>s anémones et <strong>de</strong>s<br />

gastéropo<strong>de</strong>s <strong>de</strong> gran<strong>de</strong> taille dont certains sont proches <strong>de</strong> ceux trouvés dans le pacifique<br />

occi<strong>de</strong>ntal (Hashimoto et al 2001, Van Dover et al 2001a) alors que d’autres semblent<br />

endémique <strong>de</strong> l’océan Indien (gastéropo<strong>de</strong> possédant un pied avec <strong>de</strong>s écailles, Waren et al<br />

2003, Fig. 1.13). La fermeture <strong>de</strong> la mer <strong>de</strong> Tethys (90 Ma) a pu gêner la connectivité entre<br />

cette province et la dorsale médio-Atlantique (Hessler & Lonsdale 1991). Il est cependant<br />

fort probable que <strong>de</strong>s connections récentes aient pu être établies au travers <strong>de</strong> l’Atlantique<br />

sud et le long <strong>de</strong> la dorsale Antarctique.<br />

Outre ces cinq provinces biogéographiques généralement admises, <strong>de</strong>s auteurs<br />

proposent <strong>de</strong> diviser certaines <strong>de</strong> ces provinces en sous provinces biogéographiques<br />

différentes sur la base d’analyses multivariées <strong>de</strong> la composition faunistique en<br />

présence/absence (Fig. 1.13). Van Dover et al (2002) proposent ainsi <strong>de</strong> subdiviser la<br />

dorsale médio-atlantique en <strong>de</strong>ux provinces distinctes (plateau <strong>de</strong>s Azores vs region<br />

equatoriale <strong>de</strong> la ri<strong>de</strong> médio-atlantique) en supposant que la formation du plateau <strong>de</strong>s<br />

Azores (∼20 Ma) ait pu isoler les faunes par vicariance. Bachraty et al (2009) suggèrent<br />

d’intégrer les sources hydrothermales <strong>de</strong> l’océan Indien aux bassins arrière arc les plus sud<br />

<strong>de</strong> la province du Pacifique occi<strong>de</strong>ntal et <strong>de</strong> séparer la province du Pacifique oriental en<br />

<strong>de</strong>ux provinces (nord et sud). La majorité <strong>de</strong>s espèces sont néanmoins capables <strong>de</strong> franchir<br />

cette barrière biogéographique équatoriale. Pour ces espèces, il est donc intéressant<br />

d’étudier l’impact <strong>de</strong> cette barrière à un niveau phylogéographique. L’histoire <strong>de</strong> la<br />

28


Introduction générale<br />

formation <strong>de</strong> l’EPR ainsi que ses caractéristiques géotectoniques peuvent également<br />

apporter <strong>de</strong>s informations précieuses pouvant permettre d’expliquer l’évolution <strong>de</strong>s<br />

espèces dans un cadre géodynamique.<br />

Dorsale médio-atlantique nord et sud<br />

Dorsale du Pacifique nord<br />

Ridgeia piscesae<br />

Paralvinella<br />

Rimicaris exoculata<br />

Bassins arrière arc du Pacifique occi<strong>de</strong>ntal<br />

Bathymodiolus azoricus<br />

Dorsale du Pacifique oriental<br />

Alvinella pompejana<br />

Ifremeria nautilei<br />

Océan Indien<br />

Alviniconcha hessleri<br />

Bathymodiolus brevior<br />

Riftia pachyptila<br />

Calyptogena magnifica<br />

Gastéropo<strong>de</strong> avec un<br />

pied à écailles<br />

Bathymodiolus thermophilus<br />

Fig. 1.13 Espèces emblématiques <strong>de</strong>s cinq provinces biogéographiques initialement décrites<br />

29


Introduction générale<br />

6. Dorsale du Pacifique oriental (EPR)<br />

6.1. Vitesses d’accrétion et géomorphologie du rift<br />

La morphologie <strong>de</strong>s dorsales varie en fonction (1) <strong>de</strong> la vitesse d’accrétion et (2) du<br />

taux d’approvisionnement en magma (MacDonald et al 1991). En effet, une dorsale à<br />

faible vitesse d’accrétion (10 à 40 mm/an) présente généralement une vallée au niveau du<br />

rift axial <strong>de</strong> 1 à 3 km <strong>de</strong> profon<strong>de</strong>ur (graben axial). C’est le cas <strong>de</strong> la dorsale médio<br />

Atlantique. Une dorsale à vitesse d’accrétion élevée (90 à 170 mm/an) est caractérisée par<br />

une élévation du plancher océanique <strong>de</strong> plusieurs centaines <strong>de</strong> mètres au niveau <strong>de</strong> son axe<br />

mais pas <strong>de</strong> vallée axiale (sillon <strong>de</strong> 50 mètres <strong>de</strong> large caractérisant une dorsale en dôme).<br />

Enfin, une dorsale à vitesse d’accrétion intermédiaire (40 à 90 mm/an) peut présenter un<br />

axe élevé avec une vallée généralement peu profon<strong>de</strong> suivant la quantité <strong>de</strong> magma<br />

fournie. Une forte production <strong>de</strong> magma peut néanmoins engendrer un axe élevé en dépit<br />

d’une vitesse d’accrétion faible.<br />

L’axe <strong>de</strong>s dorsales, séparant <strong>de</strong>ux plaques tectoniques, n’est pas continu. Les<br />

ruptures d’axe <strong>de</strong> la dorsale ont lieu dans les régions présentant une quantité <strong>de</strong> magma<br />

moindre (MacDonald et al 1991). Dans le cas d’accrétion rapi<strong>de</strong>, le magma se concentre en<br />

vastes réservoirs formant ainsi, sous la dorsale, un véritable collier <strong>de</strong> perles. C’est donc<br />

entre chaque perle (chaque réservoir, i.e. dans la zone <strong>de</strong> discontinuités <strong>de</strong>s processus <strong>de</strong><br />

convection thermique) qu’ont lieu les ruptures <strong>de</strong> l’axe <strong>de</strong> la dorsale. Au contraire, lorsque<br />

l’accrétion est lente, le magma ne se concentre pas en vaste réservoir mais forme <strong>de</strong><br />

multiples « gouttelettes » sous la dorsale. Ainsi, les points <strong>de</strong> rupture <strong>de</strong> l’axe <strong>de</strong> la dorsale<br />

seront plus nombreux <strong>pour</strong> une dorsale à vitesse d’accrétion lente, augmentant ainsi par la<br />

même, les barrières potentielles à la dispersion. Une dorsale à vitesse d’accrétion élevée<br />

présentera, quant à elle, une activité magmatique plus importante, pouvant augmenter la<br />

probabilité d’extinction/recolonisation <strong>de</strong>s populations hydrothermales associées.<br />

L’EPR est une dorsale à vitesse d’accrétion variable selon le segment considéré (<strong>de</strong><br />

80-125 mm/an et <strong>de</strong> 125-152 mm/an respectivement le long <strong>de</strong>s frontières Pacifique-Cocos<br />

et Pacifique-Nazca : Cormier 1997). Elle présente ainsi un <strong>de</strong>gré <strong>de</strong> segmentation moindre<br />

que la dorsale médio-atlantique (MAR, Mag<strong>de</strong> & Spark 1997). Cependant, on observe une<br />

forte diversité structurale entre les différents segments <strong>de</strong> la dorsale allant <strong>de</strong>s dorsales en<br />

30


Introduction générale<br />

dôme caractérisée par un volcanisme <strong>de</strong> type effusif (7°25S ou 17°25S avec <strong>de</strong> nombreux<br />

épanchements <strong>de</strong> lave en coussin) à <strong>de</strong>s dorsales entaillées par <strong>de</strong>s graben profonds et<br />

fortement tectonisés (21°25S)(cf Renard et al 1985). On peut ainsi penser que les<br />

populations <strong>de</strong> l’EPR sont moins structurées que celles <strong>de</strong> MAR. Cependant, la plus gran<strong>de</strong><br />

diversité biologique spécifique <strong>de</strong> l’EPR (Tunnicliffe 1991) suggère que la faune <strong>de</strong> la<br />

dorsale EPR <strong>pour</strong>rait être plus ancienne que celle <strong>de</strong> MAR, laissant plus <strong>de</strong> temps à la<br />

faune <strong>pour</strong> se différencier. La vitesse d’accrétion <strong>de</strong> l’EPR sud (entre 14°S et 22°S) est la<br />

plus élevée au mon<strong>de</strong>. Les populations <strong>de</strong> l’EPR sud <strong>pour</strong>raient donc avoir un taux <strong>de</strong><br />

renouvellement plus grand et donc <strong>de</strong>s goulots d’étranglement (bottleneck) plus fréquents.<br />

L’histoire <strong>de</strong> la formation <strong>de</strong> l’EPR ainsi que ses caractéristiques géotectoniques<br />

sont donc autant <strong>de</strong> paramètres à prendre en considération <strong>pour</strong> expliquer les patrons<br />

phylogéographiques <strong>de</strong>s espèces <strong>de</strong> cette dorsale.<br />

6.2. Définition et principe d’une faille transformant et d’une<br />

microplaque<br />

En modifiant l’axe <strong>de</strong> la vallée axiale <strong>de</strong> la dorsale, les failles transformantes peuvent<br />

engendrer <strong>de</strong>s barrières à la dispersion. Ces failles sont la conséquence <strong>de</strong>s forces <strong>de</strong><br />

cisaillement portées en différents points <strong>de</strong> la dorsale selon les vitesses d’accrétion et les<br />

mouvements <strong>de</strong>s plaques. Ces failles telles que celles <strong>de</strong> la zone <strong>de</strong> fracture Rivera peuvent<br />

décaler la dorsale sur plusieurs centaines <strong>de</strong> kilomètres et constituer <strong>de</strong>s disruptions dans<br />

l’écoulement <strong>de</strong>s courants <strong>de</strong> fonds à l’axe <strong>de</strong>s dorsales. En général, ces failles datent <strong>de</strong><br />

quelques millions d’années (1 à 4 Ma).<br />

La formation d’une microplaque peut se résumer en trois étapes (Fig. 1.14). La<br />

première consiste en une fracturation <strong>de</strong> la dorsale en <strong>de</strong>ux morceaux (rupture <strong>de</strong> l’axe par<br />

<strong>de</strong>s failles transformantes). La secon<strong>de</strong> étape est la propagation <strong>de</strong>s <strong>de</strong>ux morceaux <strong>de</strong><br />

dorsale en sens opposé, constituant ainsi un système <strong>de</strong> <strong>de</strong>ux morceaux <strong>de</strong> dorsale<br />

parallèles. Dans la troisième étape, les <strong>de</strong>ux morceaux <strong>de</strong> dorsale se rejoignent <strong>pour</strong> former<br />

une microplaque.<br />

31


Introduction générale<br />

A B C D<br />

3<br />

1<br />

M<br />

2<br />

4<br />

Fig. 1.14 Schématisation <strong>de</strong> la formation d’une microplaque.<br />

A, dorsale sans fracturation ; B, fracturation <strong>de</strong> la dorsale ; C, propagation en sens inverse<br />

<strong>de</strong>s morceaux <strong>de</strong> dorsale ; D, isolement <strong>de</strong> la microplaque (M).<br />

1, Ouest <strong>de</strong> la microplaque ; 2, Est <strong>de</strong> la microplaque; 3, jonction nord ; 4, jonction sud.<br />

6.3. Histoire <strong>de</strong> la formation <strong>de</strong> la dorsale EPR<br />

La dorsale du Pacifique oriental telle qu’on la considère actuellement (EPR) est située<br />

à l’ouest <strong>de</strong> l’Amérique du Sud entre les plaques Pacifique et Cocos au Nord et entre les<br />

plaques Pacifique et Nazca au Sud. Elle s’étend ainsi entre les latitu<strong>de</strong>s 30°N et 40°S. La<br />

dislocation <strong>de</strong> la plaque Farallon (par subduction sous la plaque Pacifique) a produit un<br />

plancher océanique structurellement chaotique entraînant la naissance <strong>de</strong> la microplaque<br />

Mendoza, il y a 24 Ma (Eakins & Lonsdale 2003, Fig. 1.15).<br />

32


Introduction générale<br />

Fig. 1.15 Schématisation <strong>de</strong> l’histoire évolutive <strong>de</strong> l’EPR.<br />

En blanc : dorsale nord pacifique<br />

En rouge : dorsale ancestrale qui s’inactive (pointillés) ayant donné lieu à l’EPR actuelle, à<br />

partir du point <strong>de</strong> fracture (cercle noir). Les flèches indiquent la propagation <strong>de</strong>s dorsales.<br />

MR, Mathematician Ridge ; MT, Moctezuma Trough, GR, Galapagos Rise<br />

Cercle vi<strong>de</strong> : microplaque. mr, microplaque mathématicienne ; mb, microplaque Bauer ; mp,<br />

microplaque <strong>de</strong> Pâques ; mjf, microplaque Juan Fernan<strong>de</strong>z<br />

33


Introduction générale<br />

À 20-18,5 Ma, à 13°S, la zone <strong>de</strong> fracture Marquesas/Mendana donne « naissance » à une<br />

autre zone <strong>de</strong> fracture appelée l’oEPR (EPR d’origine, Fig. 1.15). L'ancienne dorsale<br />

s’inactive progressivement et se trouve séparée en plusieurs morceaux fossiles: Selkirk<br />

Rise, Roggeveen Rise, Mendoza Rise et Galápagos Rise à l'Est <strong>de</strong> l'oEPR et Gallego Rise à<br />

l'Ouest (Mammerick et al 1980). L’oEPR connaît ensuite différentes modifications avec la<br />

formation et/ou capture <strong>de</strong> diverses microplaques.<br />

6.3.1. La microplaque Bauer (Eakins & Lonsdale 2003) : 10°S-18°S<br />

La microplaque <strong>de</strong> Bauer a commencé sa formation il y a 17 Ma à partir d’un<br />

dédoublement du centre d’accrétion <strong>de</strong> l’oEPR. Les <strong>de</strong>ux morceaux <strong>de</strong> dorsale générés par<br />

faille transformante sont alors nommés iEPR <strong>pour</strong> la fraction sud (1, Fig. 1.14 et la<br />

Galápagos Rise (GR, ne correspond pas à la dorsale Cocos-Nazca située au large <strong>de</strong>s îles<br />

Galápagos) <strong>pour</strong> la fraction nord (2, Fig. 1.14). De 15 à 11 Ma, les dorsales iEPR et GR se<br />

propagent l’une vers l’autre. Les rifts formés à l’extrémité <strong>de</strong>s axes <strong>de</strong> propagation<br />

engendrent un relief escarpé à la marge <strong>de</strong> ces dorsales. Cette propagation <strong>de</strong>s dorsales<br />

s’effectue <strong>de</strong> manière oblique par rapport à leur pôle <strong>de</strong> rotation respectif, engendrant ainsi<br />

la formation <strong>de</strong> nouvelles fractures (offsets) perpendiculaires aux axes d’accrétion. À 11<br />

Ma, iEPR et Galapagos Rise se remettent en contact par l’intermédiaire <strong>de</strong>s failles<br />

transformantes Garrett (3, Fig. 1.14) et Dana et Woce (4, Fig. 1.14), isolant ainsi la<br />

microplaque Bauer.<br />

6.3.2. La paléoplaque Mathematician (Mammerick et al 1980, Mammerick &<br />

Klitgord 1982) : 12°N-17°N<br />

À 10 Ma, la Moctozuma Trough se forme à l’Est <strong>de</strong> la Mathematician Ridge entre<br />

12°N et 17°N (Mammerick & Klitgord 1982, Fig. 1.15), parallèlement à celle-ci.<br />

Moctozuma Trough et Mathematician Ridge se rejoignent (formation <strong>de</strong>s zones <strong>de</strong> fracture<br />

Riviera et O’Gorman au Nord et au Sud, respectivement), formant la paléoplaque<br />

Mathematician. Elles coexistent pendant 3-5 Ma puis la Mathematician Ridge s’inactive<br />

(7-5 Ma), engendrant l’annexion <strong>de</strong> la paléoplaque Mathematician à la plaque Pacifique<br />

(Mammerick & Klitgord 1982). La Moctezuma Trough est ainsi rattachée à l’iEPR <strong>pour</strong><br />

donner l’EPR actuelle (Mammerick et al 1980).<br />

34


Introduction générale<br />

6.3.3. Capture <strong>de</strong> la microplaque Bauer et formation <strong>de</strong>s microplaques <strong>de</strong><br />

Pâques (Naar & Hey 1991, Rusby & Searle 1995, ~25°S) et Juan<br />

Fernan<strong>de</strong>z (Larson et al 1992, 30°S-36°S)<br />

À ~5,8 Ma, la microplaque Bauer (10°S-18°S) est capturée par subduction, sous la<br />

plaque Nazca (Eakins & Lonsdale 2003) et n’existe donc plus actuellement.<br />

À peu près à la même pério<strong>de</strong> que la capture <strong>de</strong> la microplaque Bauer, ~5,89 Ma,<br />

l’EPR se fracture à la latitu<strong>de</strong> d’environ 25°S isolant 2 segments <strong>de</strong> dorsales (Naar & Hey<br />

1991, Rusby & Searle 1995). À 5,25 Ma, le segment <strong>de</strong> dorsale situé au Sud Est <strong>de</strong> cette<br />

fracture commence sa propagation vers le Nord. À 2,5 Ma, sa vitesse <strong>de</strong> propagation<br />

diminue et s’arc-boute sur le segment <strong>de</strong> dorsale situé au Nord Ouest, lequel entame une<br />

propagation en forme d’arc vers le sud, imprimant un début <strong>de</strong> rotation <strong>de</strong> la microplaque<br />

<strong>de</strong> Pâques dans le sens inverse <strong>de</strong>s aiguilles d’une montre (contrairement à la microplaque<br />

<strong>de</strong> Bauer et à la schématisation sur la Fig. 1.14). La dorsale occi<strong>de</strong>ntale isole la<br />

microplaque au Sud et la dorsale orientale cesse sa propagation. Actuellement, la<br />

microplaque <strong>de</strong> Pâques n’est toujours pas isolée au Nord (Fig. 1.16).<br />

Encore un peu plus au sud, la microplaque Juan Fernan<strong>de</strong>z (Fig. 1.16, Larson et al<br />

1992) est issue <strong>de</strong> la propagation vers le Sud <strong>de</strong> ce qui reste du segment sud <strong>de</strong> la dorsale<br />

EPR ayant joué un rôle dans la formation <strong>de</strong> la microplaque <strong>de</strong> l’île <strong>de</strong> Pâques, et sa<br />

rencontre, il y a 3,4 Ma, avec la dorsale Pacifique-Antarctique s’étendant vers le nord. À<br />

2,5 Ma, la microplaque est isolée au Sud puis s’isole complètement au Nord (0,7 Ma) via<br />

la Deep En<strong>de</strong>avour.<br />

35


Introduction générale<br />

Fig. 1.16 Géomorphologie actuelle <strong>de</strong> la dorsale pacifique orientale.<br />

PAR, Dorsale Pacifique-Antarctique ; ZF, zone <strong>de</strong> fracture<br />

6.4. Potentielles barrières à la dispersion<br />

6.4.1. Failles transformantes<br />

Les failles transformantes retrouvées à l’équateur sont datées <strong>de</strong> 1-2 Ma entre 9°N et<br />

17°S (Kureth & Rea 1981, Naar & Hey 1989, Francheteau et al 1990, Fig. 1.16). L’impact<br />

<strong>de</strong> ces failles sur les espèces en tant que barrière à la dispersion peut différer en fonction<br />

36


Introduction générale<br />

<strong>de</strong>s capacités <strong>de</strong> dispersion <strong>de</strong> ces espèces et du comportement <strong>de</strong> leurs larves potentielles<br />

dans la colonne d’eau. Ainsi, Johnson et al (2006) ont montré que la faille transformante<br />

Blanco (dorsale pacifique nord) datée <strong>de</strong> 3,5 Ma serait responsable <strong>de</strong> la spéciation<br />

allopatrique <strong>de</strong> Lepetodrilus fucensis et L. gor<strong>de</strong>nsis alors que son impact est moindre<br />

<strong>pour</strong> Ridgeia piscisae (différenciation génétique mais absence <strong>de</strong> monophylie réciproque,<br />

Young et al 2008).<br />

6.4.2. Courants océaniques profonds<br />

Certains courants océaniques profonds transverse à l’axe <strong>de</strong> la dorsale peuvent<br />

également réduire la dispersion et jouer un rôle <strong>de</strong> barrière. Ainsi, divers gyres ont été mis<br />

en évi<strong>de</strong>nce (Reid 1997) le long <strong>de</strong> l’EPR. Au niveau <strong>de</strong>s zones <strong>de</strong> contact entre 2 gyres, la<br />

dispersion latitudinale <strong>de</strong>s larves <strong>pour</strong>rait ainsi être limitée. Ces zones <strong>de</strong> contact sont<br />

situées au niveau <strong>de</strong> l’équateur, aux environs <strong>de</strong> 13°N, 15°S et <strong>de</strong> 32°S (Fig. 1.17).<br />

Fig. 1.17 Courants océaniques profonds (en rouge et bleu) proche <strong>de</strong> l’EPR et <strong>de</strong> la<br />

dorsale Cocos-Nazca (en jaune). Modifiée <strong>de</strong> Reid 1997.<br />

37


Introduction générale<br />

7. Structure génétique <strong>de</strong>s espèces <strong>de</strong> la dorsale du Pacifique oriental<br />

Les caractéristiques éphémères et fragmentées du milieu hydrothermal ont<br />

rapi<strong>de</strong>ment suscité <strong>de</strong> nombreuses interrogations quant aux capacités <strong>de</strong> dispersion <strong>de</strong>s<br />

espèces et au <strong>de</strong>gré <strong>de</strong> connexion entre les populations. Par exemple, <strong>de</strong>s étu<strong>de</strong>s menées<br />

sur le bivalve Calyptogena magnifica ont montré que, bien que présentant <strong>de</strong> la structure<br />

génétique entre population (Hurtado et al 2003), les larves <strong>de</strong> cette espèce seraient<br />

capables <strong>de</strong> se disperser sur l’ensemble <strong>de</strong> leur aire <strong>de</strong> distribution (i.e. plusieur milliers <strong>de</strong><br />

km, Karl et al 1996). De plus, Kojima et al (2004) ont mis en détecter au moins huit<br />

événements <strong>de</strong> migration tras-Pacifique, <strong>de</strong>ux évènements entre le Pacifique Nord-Ouest et<br />

Sud-Ouest et <strong>de</strong>ux autres les océans Pacifique et Atlantique dans l’histoire évolutive <strong>de</strong>s<br />

Vesicomidae.<br />

L’EPR suscite également un intérêt d’autant plus grand qu’elle présente <strong>de</strong>s vitesses<br />

d’accrétion importantes et donc un fort taux d’extinction/recolonisation. Des approches<br />

directes visant à comprendre le comportement <strong>de</strong>s larves ont ainsi été réalisées (e.g.<br />

recrutement <strong>de</strong>s larves sur <strong>de</strong>s blocs <strong>de</strong> basalte déployés sur le site 9°50’N, Mullineaux et<br />

al 1998 ; processus <strong>de</strong> développement <strong>de</strong>s embryons d’A. pompejana en fonction <strong>de</strong>s<br />

conditions environementales, Pradillon et al 2005) mais restent relativement difficiles à<br />

mettre en œuvre du fait <strong>de</strong> la faible accessibilité du milieu. Des approches indirectes,<br />

encore peu nombreuses, basées sur l’analyse <strong>de</strong> la structure génétique <strong>de</strong>s espèces<br />

hydrothermales ont ainsi été initiées afin <strong>de</strong> mieux comprendre l’impact <strong>de</strong> la dispersion<br />

sur la distribution spatiale <strong>de</strong>s espèces ainsi que l’impact <strong>de</strong> l’environnement (en termes <strong>de</strong><br />

fragmentation et d’instabilité) sur le <strong>de</strong>gré <strong>de</strong> connectivité (migration efficace) et les<br />

fluctuations en taille <strong>de</strong>s populations<br />

7.1. Quels modèles <strong>de</strong> dispersion <strong>pour</strong> les espèces <strong>de</strong> l’EPR ?<br />

Sur la base <strong>de</strong>s données génétiques, les espèces hydrothermales étudiées sur l’EPR<br />

semblent capables <strong>de</strong> disperser sur <strong>de</strong>s centaines <strong>de</strong> kilomètres et répondraient à un modèle<br />

classique <strong>de</strong> métapopulation <strong>de</strong> Levin (modèle en îles + extinction/recolonisation). Ces<br />

données apparaissent en contradiction avec l’hypothèse première d’un isolement linéaire<br />

par la distance proposée par Lutz (1988) sur la base d’une distribution fragmentée et uni-<br />

38


Introduction générale<br />

directionnelle <strong>de</strong> l’habitat hydrothermal et du mo<strong>de</strong> <strong>de</strong> développement lécithotrophique <strong>de</strong>s<br />

larves prépondérant en milieu hydrothermal (cf Tyler & Young 1999). Bien que ne<br />

présentant pas les mêmes types <strong>de</strong> larves (planctotrophes versus lécithotrophes), <strong>de</strong>s étu<strong>de</strong>s<br />

utilisant <strong>de</strong>s marqueurs allozymiques et/ou <strong>de</strong>s séquences du gène mitochondrial<br />

Cytochrome Oxydase I (COI) ont ainsi montré que les structures génétiques <strong>de</strong><br />

Bathymodiolus thermophilus (Grassle 1985, Craddock et al 1995) et Alvinella pompejana<br />

(Jollivet et al 1995) se conformeraient à un modèle en îles (migration importante entre les<br />

sites). En modélisant les capacités <strong>de</strong> dispersion <strong>de</strong> ces espèces, Jollivet et al (1999)<br />

proposent d’ajuster la dispersion <strong>de</strong> B. thermophilus à un modèle <strong>de</strong> métapopulation <strong>de</strong><br />

type « migrant ». Pareillement, les étu<strong>de</strong>s allozymiques effectuées sur les gastéropo<strong>de</strong>s<br />

Lepetodrilus pustulosus et Eulepetopsis vitrea (Craddock et al 1997) n’ont pas montré<br />

d’isolement par la distance et, suggèrent également une dispersion suivant un modèle en<br />

îles. À l’inverse et paradoxalement, France et al (1992) furent les premiers à proposer la<br />

présence d’un isolement par la distance chez une espèce hydrothermale vagile,<br />

l’amphipo<strong>de</strong> Ventiella sulfuris, capable <strong>de</strong> se déplacer en essaims d’un site à un autre. Ces<br />

auteurs relient néanmoins cette structure <strong>de</strong> populations au fait que l’animal a un<br />

développement direct, les oeufs fécondés se développant dans un organe appelé le<br />

marsupium. Néanmoins, la mise en évi<strong>de</strong>nce <strong>de</strong> ce modèle d’isolement apparaît biaisée par<br />

un échantillonnage <strong>de</strong> <strong>de</strong>ux groupes <strong>de</strong> sous-populations différenciées fortement distantes<br />

respectivement sur l’EPR nord et la dorsale Cocos-Nazca. L’hypothèse d’une telle barrière<br />

entre les 2 dorsales avait d’ailleurs été émise par Grassle (1985) et Jollivet et al (1995)<br />

<strong>pour</strong> expliquer l’augmentation <strong>de</strong> la différenciation génétique observée chez les espèces<br />

Bathymodiolus thermophilus et Paralvinella grasslei. De nouvelles analyses effectuées<br />

uniquement avec les sites <strong>de</strong> l’EPR (Vrijenhoek 1997) ont abouti, sous l’hypothèse du<br />

modèle en îles, à <strong>de</strong> fortes valeurs <strong>de</strong> migration entre les sites <strong>de</strong> l’EPR (11,6 individus par<br />

génération). Si un isolement par la distance a également été détecté à cette échelle spatiale,<br />

il semblerait qu’il s’agisse également d’un second biais résultant <strong>de</strong> la séparation <strong>de</strong>s<br />

populations <strong>de</strong> part et d’autre <strong>de</strong> la zone <strong>de</strong> fracture Riviera (Vrijenhoek 1997). Dans ce<br />

cas, la différenciation observée résulte plus <strong>de</strong> la mise en place <strong>de</strong> barrières physiques à la<br />

dispersion qu’à un effet <strong>de</strong> la distance géographique. Des échantillonnages et analyses<br />

complémentaires s’avèrent donc nécessaires <strong>pour</strong> infirmer ou confirmer l’existence d’un<br />

isolement par la distance chez V. sulfuris. Au contraire, la structure génétique du<br />

siboglinidae Riftia pachyptila semble, quant à elle, s’ajuster <strong>de</strong> manière moins critiquable à<br />

un modèle avec isolement par la distance (Black et al 1994), et ce malgré une autonomie<br />

39


Introduction générale<br />

dispersive <strong>de</strong> la larve assez gran<strong>de</strong> (3 semaines à 1 mois : Marsh et al 2001). L’étu<strong>de</strong>,<br />

réalisée sur un nombre <strong>de</strong> champs hydrothermaux plus importants, prend en effet soin <strong>de</strong><br />

supprimer les sites ayant potentiellement divergé <strong>de</strong> part et d’autre d’une barrière à la<br />

dispersion (Vrijenhoek 1997). À l’heure actuelle, R. pachyptila est la seule espèce étudiée<br />

<strong>de</strong> l’EPR <strong>pour</strong> laquelle un isolement par la distance a été mis en évi<strong>de</strong>nce. Cependant,<br />

Shank & Halanych (2007) ont toutefois récemment montré l’existence <strong>de</strong> migrants sur <strong>de</strong><br />

longues distances entre 9°50’N et Guyamas chez cette espèce, allant à l’encontre <strong>de</strong>s<br />

résultats <strong>de</strong> Black et al (1994).<br />

7.2. Existence <strong>de</strong> barrières génétiques chez les espèces <strong>de</strong> l’EPR<br />

Les premières étu<strong>de</strong>s génétiques sur l’EPR ne se sont intéressées qu’à une échelle<br />

géographique restreinte (EPR nord) sans pouvoir mettre en évi<strong>de</strong>nce la présence <strong>de</strong><br />

barrière à la dispersion à cette échelle (A. pompejana, Jollivet et al 1995 ; B. thermophilus,<br />

Craddock et al 1995) et, ce malgré la présence <strong>de</strong> la FZ Rivera (la faille la plus longue<br />

connue sur l’EPR). Néanmoins, la plupart <strong>de</strong> ces étu<strong>de</strong>s démontrent une différenciation <strong>de</strong>s<br />

populations entre l’EPR nord et la dorsale Cocos-Nazca (B. thermophilus : Grassle 1985 ;<br />

V. sulfuris : France et al 1992 ; Paralvinella grasslei : Jollivet et al 1995 ; Calyptogena<br />

magnifica : Karl et al 1996).<br />

Les étu<strong>de</strong>s plus récentes basées sur l’analyse <strong>de</strong> séquences ADN ont élargi le champ<br />

d’investigation en échantillonnant <strong>de</strong>s sites provenant <strong>de</strong> l’EPR sud, et révélé l’existence<br />

<strong>de</strong> diverses barrières génétiques par <strong>de</strong>s métho<strong>de</strong>s plus sensibles utilisant la divergence<br />

inter-individuelle :<br />

(1) Des cla<strong>de</strong>s monophylétiques ont été i<strong>de</strong>ntifiés chez A. pompejana <strong>de</strong> part et<br />

d’autre <strong>de</strong> l’équateur, suggérant un impact important d’une barrière située à ces<br />

latitu<strong>de</strong>s, qu’elle soit <strong>de</strong> type hydrologique gyre ou <strong>de</strong> nature tectonique (Hess<br />

Deep, failles transformantes Quebrada/Discovery/Gofar). Cette barrière <strong>pour</strong>rait<br />

également, dans une moindre mesure, être responsable <strong>de</strong> la différenciation <strong>de</strong>s<br />

populations <strong>de</strong> Tevnia jerichonana.<br />

(2) La microplaque <strong>de</strong> Pâques aurait joué un rôle <strong>de</strong> barrière à la dispersion chez R.<br />

pachyptila (COI : Hurtado et al 2004) mais également chez B. thermophilus<br />

(COI et allozymes : Won et al 2003). Won et al (2003) suggère ainsi que la<br />

40


Introduction générale<br />

divergence accumulée <strong>de</strong> part et d’autre <strong>de</strong> cette microplaque a engendré <strong>de</strong>ux<br />

espèces cryptiques chez B. thermophilus par spéciation allopatrique. Les<br />

polychètes Branchipolynoe symmytilida et A. pompejana ne semblent cependant<br />

pas affectés par cette barrière (Hurtado et al 2004, COI).<br />

(3) Won et al (2003) ont montré l’existence d’une différenciation génétique sur B.<br />

thermophilus située entre 11°S et 17°S. Ces auteurs ont suggéré que cette<br />

différenciation génétique <strong>pour</strong>rait être due à un gyre profond présent à 15°S.<br />

(4) Une structure génétique a également été observée entre plusieurs populations <strong>de</strong><br />

l’EPR chez Riftia pachyptila via <strong>de</strong>s AFLP (9°N/17°S, Shank & Halanych<br />

2007). De même, <strong>de</strong>ux espèces cryptiques avec potentialités d’hybridation ont<br />

été mises en évi<strong>de</strong>nce chez Lepetodrilus elevatus, la première échantillonnée à<br />

13°N et 9°50’N et la secon<strong>de</strong> à 9°50’N et 17°S (COI et allozymes, Matabos et<br />

al 2008b). Néanmoins, <strong>de</strong>s analyses complémentaires sont nécessaires <strong>pour</strong><br />

définir plus précisément la localisation <strong>de</strong>s barrières potentielles aux flux <strong>de</strong><br />

gènes chez ces espèces.<br />

Ces étu<strong>de</strong>s constituent ainsi les prémices d’une étu<strong>de</strong> <strong>de</strong> phylogéographie comparée<br />

sur la dorsale du Pacifique oriental. Peu d’analyses (cf. Hurtado et al 2004) ont <strong>pour</strong> le<br />

moment été effectuées afin <strong>de</strong> tester ces hypothèses d’événements <strong>de</strong> vicariance communs<br />

entre espèces, ou d’évaluer le sens <strong>de</strong> colonisation préférentiel <strong>de</strong> la dorsale. Si certaines<br />

étu<strong>de</strong>s ont tenté <strong>de</strong> détecter <strong>de</strong> l’hybridation entre les lignées divergentes observées sur le<br />

COI en couplant ce gène à <strong>de</strong>s données allozymiques (Karl et al 1996, Won et al 2003,<br />

Matabos et al 2008b), le pouvoir résolutif <strong>de</strong>s allozymes reste toutefois faible comparé à<br />

celui <strong>de</strong>s séquences <strong>de</strong> gènes nucléaires. L’utilisation <strong>de</strong> gènes nucléaires <strong>pour</strong>rait ainsi<br />

permettre <strong>de</strong> mieux évaluer les contours d’une espèce donnée et notamment <strong>de</strong> savoir si<br />

l’existence <strong>de</strong> lignées mitochondriales distinctes coinci<strong>de</strong> réellement avec l’existence<br />

d’espèces cryptiques (notion <strong>de</strong> monophilies réciproques à tous les gènes) ou, a contrario<br />

<strong>de</strong> comprendre ces phénomènes d’hybridation potentiels par l’utilisation <strong>de</strong> métho<strong>de</strong>s<br />

basées sur la théorie <strong>de</strong> la coalescence. De plus, la comparaison <strong>de</strong> plusieurs gènes et/ou<br />

plusieurs espèces <strong>pour</strong>rait permettre <strong>de</strong> mieux appréhen<strong>de</strong>r l’histoire démographique <strong>de</strong>s<br />

populations <strong>de</strong> l’EPR, jusqu’alors largement négligée et notamment son rôle dans la<br />

structuration génétique <strong>de</strong>s populations actuelles. Enfin, l’approche multi-locus fournit un<br />

moyen particulièrement puissant <strong>de</strong> détecter <strong>de</strong>s effets sélectifs potentiels sur certains<br />

gènes.<br />

41


Introduction générale<br />

8. Objectifs <strong>de</strong> la thèse<br />

L’objectif <strong>de</strong> ce travail est <strong>de</strong> réaliser une étu<strong>de</strong> phylogéographique <strong>de</strong> la dorsale du<br />

Pacifique oriental en combinant <strong>de</strong>s approches classiques d’analyses <strong>de</strong> fréquences<br />

alléliques dans les populations et <strong>de</strong> leurs distributions géographiques avec <strong>de</strong>s métho<strong>de</strong>s<br />

d’analyse plus récentes basées sur la théorie <strong>de</strong> la coalescence. Ces approches combinées<br />

permettent <strong>de</strong> détecter d’éventuelles barrières à la dispersion, <strong>de</strong> les localiser spatialement,<br />

et <strong>de</strong> mesurer leur <strong>de</strong>gré <strong>de</strong> perméabilité. Les étu<strong>de</strong>s <strong>de</strong> coalescence permettent <strong>de</strong> plus<br />

d’en évaluer l’âge, <strong>de</strong> chercher à savoir s’il existe un sens <strong>de</strong> la colonisation le long <strong>de</strong> la<br />

dorsale et <strong>de</strong> mieux comprendre l’histoire démographique <strong>de</strong>s espèces en présence. Pour<br />

cela, nous avons d’abord opté <strong>pour</strong> une approche comparative entre sept espèces en<br />

regardant plus spécifiquement le gène mitochondrial COI. Nous avons ensuite traité<br />

plusieurs cas d’espèces à l’ai<strong>de</strong> <strong>de</strong> plusieurs types <strong>de</strong> marqueurs (COI, gènes nucléaires,<br />

allozymes) en choisissant <strong>de</strong>s espèces appartenant à 2 embranchements : annéli<strong>de</strong> et<br />

mollusque. Les 3 espèces étudiées (le ver <strong>de</strong> pompei Alvinella pompejana, le bivalve<br />

Bathymodiolus thermophilus et le gastéropo<strong>de</strong> patelliforme Lepetodrilus elevatus) ont été<br />

choisies <strong>pour</strong> 2 raisons : (1) <strong>de</strong>s stratégies reproductives et <strong>de</strong>s mo<strong>de</strong>s <strong>de</strong> développement<br />

larvaire différents et (2) il s’agit d’espèces <strong>pour</strong> lesquelles <strong>de</strong>s banques <strong>de</strong> gènes sont<br />

disponibles. L’approche multi espèces avait notamment <strong>pour</strong> but <strong>de</strong> tester <strong>de</strong>s hypothèses<br />

d’événements <strong>de</strong> vicariance associés à une éventuelle séparation <strong>de</strong>s faunes et chercher à<br />

savoir si ces espèces pouvaient avoir eu une histoire démographique commune.<br />

L’approche multi gènes nous a permis, quant à elle, d’appréhen<strong>de</strong>r les phénomènes<br />

d’hybridation entre lignées divergentes, <strong>de</strong> mieux comprendre les patrons <strong>de</strong> migration et<br />

<strong>de</strong> détecter <strong>de</strong>s gènes potentiellement sous sélection.<br />

42


Chapitre 2<br />

Chapitre 2 : Métho<strong>de</strong>s et biais méthodologiques<br />

1. Stratégie d’échantillonnage<br />

Les échantillons utilisés <strong>pour</strong> ce travail <strong>de</strong> thèse sont principalement issus <strong>de</strong><br />

plusieurs campagnes océanographiques françaises le long <strong>de</strong> la dorsale EPR. Les premières<br />

campagnes (HOT96, HOPE99, PHARE2002) ont été menées au Nord <strong>de</strong> l’équateur sur les<br />

sites situés à 13°N et 9°50’N et avaient notamment <strong>pour</strong> objectif d’étudier les flux<br />

géniques entre les sites hydrothermaux présents au sein d’un ou <strong>de</strong>ux champs<br />

hydrothermaux (un champ hydrothermal étant composé <strong>de</strong> plusieurs sites). La campagne<br />

océanographique BIOSPEEDO 2004 visait à élargir l’étu<strong>de</strong> <strong>de</strong>s flux <strong>de</strong> gènes à <strong>de</strong>s champs<br />

hydrothermaux situés au Sud <strong>de</strong> l’EPR afin <strong>de</strong> mieux comprendre le rôle <strong>de</strong>s failles<br />

transformantes, microplaques et courants océaniques profonds dans la structure génétique<br />

<strong>de</strong>s populations. Une collaboration avec Timothy Mark Shank, chercheur au Woods Hole<br />

Oceanographic Institute a permis <strong>de</strong> compléter cet échantillonnage au Nord <strong>de</strong> la dorsale<br />

(campagne Extreme 2003) <strong>pour</strong> l’espèce L. ovalis qui n’avait pas été récoltée lors <strong>de</strong>s<br />

campagnes françaises, l’espèce n’ayant pas été trouvée.<br />

Il est nécessaire <strong>de</strong> prendre conscience <strong>de</strong> la difficulté d’échantillonner en milieu<br />

hydrothermal profond. En effet, les sites étant généralement situé à plus <strong>de</strong> 2000 m <strong>de</strong><br />

profon<strong>de</strong>ur, l’accès aux échantillons nécessite <strong>de</strong>s moyens technologiques et humains<br />

importants. Les prélèvements sont ainsi effectués par <strong>de</strong>s submersibles hautement<br />

résistants aux fortes pressions (sphère titane). Ces submersibles peuvent être habités (<strong>pour</strong><br />

cette étu<strong>de</strong>, utilisation du sous-marin français Nautile lors <strong>de</strong>s campagnes HOT96,<br />

HOPE99 et BIOSPEEDO 2004, et du sous-marin américain Alvin lors <strong>de</strong> la campagne<br />

Extreme 2003, Fig. 2.1.A, 2.1.B, 2.1.C) ou télécommandé <strong>de</strong>puis la surface à partir d’une<br />

laisse avec un câble électroporteur (<strong>pour</strong> l’alimentation en énergie) et <strong>de</strong>s faisceaux <strong>de</strong><br />

fibres optiques (<strong>pour</strong> récupérer les informations) (utilisation du Remote Operated<br />

Vehicule, ROV Victor 6000 <strong>pour</strong> la campagne PHARE 2002 Fig. 2.1.D). Il convient<br />

également <strong>de</strong> rappeler la difficulté d’observation liée à l’obscurité (l’éclairage <strong>de</strong>s<br />

43


Chapitre 2<br />

bathyscaphes permettant un champ <strong>de</strong> vision <strong>de</strong> quelques mètres) et le temps<br />

d’échantillonnage limité (une dizaine d’heures <strong>pour</strong> un submersible habité par plongée).<br />

Les animaux sont récoltés à l’ai<strong>de</strong> <strong>de</strong>s bras télé-manipulés <strong>de</strong> ces bathyscaphes (Fig.<br />

2.1.E), placés dans <strong>de</strong>s paniers isothermes (Fig. 2.1.F) et remontés en surface<br />

(décompression <strong>de</strong>s organismes). À bord du navire océanographique (N/O L’Atalante <strong>pour</strong><br />

les campagnes françaises et R/V Atlantis <strong>pour</strong> la campagne américaine, Fig 2.1.G, 2.1.H),<br />

ils sont alors triés par espèce et préservés dans l’éthanol absolu ou congelé à -80°C ou dans<br />

l’azote liqui<strong>de</strong> <strong>pour</strong> les extractions d’ADN/ARN réalisées à terre.<br />

44


Chapitre 2<br />

A<br />

B<br />

C<br />

D<br />

E<br />

F<br />

G<br />

H<br />

Fig. 2.1 Moyens à la mer et équipements utilisés <strong>pour</strong> effectuer les échantillonnages<br />

nécessaires à la thèse.<br />

A et B, submersible habité Nautile ; C, submersible habité Alvin ; D, ROV Victor 6000<br />

C, pince télé-manipulée (ici <strong>de</strong> l’Alvin) ; D, panier isotherme du Nautile<br />

G, le navire français N/O L’Atalante ; H, le navire américain R/V Atlantis<br />

45


Chapitre 2<br />

Ces différentes campagnes océanographiques ont ainsi permis <strong>de</strong> récolter <strong>de</strong>s<br />

populations d’espèces différentes sur <strong>de</strong>s champs hydrothermaux distribués <strong>de</strong> 21°N à<br />

21°33’S (Fig. 2.2), permettant ainsi d’effectuer les analyses <strong>de</strong> phylogéographie comparée<br />

réalisées lors <strong>de</strong> ma thèse.<br />

21°N<br />

120°W 110° 100° 90° 80°W<br />

13°N<br />

20°N<br />

9°50’N<br />

7°25’S<br />

14°S<br />

17°25’S<br />

17°34’S<br />

18°25’S<br />

18°33’S<br />

21°25’S<br />

21°33’S<br />

1000 km<br />

10°<br />

0<br />

10°<br />

20°<br />

30°S<br />

Fig. 2.2 Champs hydrothermaux échantillonnés le long <strong>de</strong> la dorsale EPR<br />

2. Banques d’ADN complémentaires<br />

Les gènes nucléaires utilisés dans le chapitre 3 sont issus <strong>de</strong> <strong>de</strong>ux banques d’ADNc<br />

réalisées par le Dr Arnaud Tanguy sur plusieurs individus <strong>de</strong> B. azoricus récoltés sur le site<br />

Menez Gwen (dorsale Médio-Atlantique, 37°50N, Tanguy et al 2008) et <strong>de</strong> L. elevatus du<br />

site Oasis (EPR, 17°25’S), respectivement. Les 2 banques effectuées sur individus entiers<br />

ont été obtenues par utilisation du kit <strong>de</strong> synthèse d’ADNc SMART TM et du kit <strong>de</strong> clonage<br />

SMART TM (Clontech) après purification <strong>de</strong>s messagers sur résine polyAT tract®mRNA.<br />

Les gènes nucléaires obtenus sur A. pompejana proviennent <strong>de</strong> banques ADNc tissulaires<br />

(branchie, tissu ventral et pygidium) réalisées par le Génoscope d’Evry dans le cadre du<br />

Consortium ‘Alvinella’ à partir <strong>de</strong> 2 types <strong>de</strong> métho<strong>de</strong>s <strong>de</strong> clonage : (1) ‘Oligocapping’ <strong>de</strong>s<br />

ARNm par phosphatase alcaline, récupération <strong>de</strong>s messagers sur résine OligodT après<br />

élimination <strong>de</strong> la coiffe 5’ et insertion dans un vecteur pME18S-FL3 après RT-PCR et, (2)<br />

46


Chapitre 2<br />

la métho<strong>de</strong> ‘CloneMiner’ ou les ADNc double brins sont produits par ajout <strong>de</strong> séquences<br />

AttB à chaque bout et clonés dans un vecteur Gateway par recombinaison homologue. Ces<br />

banques représentent un total <strong>de</strong> 100 200 séquences assemblées et annotées (Gagnière et<br />

al soumis).<br />

L’extraction <strong>de</strong>s ARN totaux a été effectuée à partir d’un protocole utilisant le Tri-<br />

Reagent. Cinquante à cent milligrammes <strong>de</strong> tissu congelé sont placés dans 1 mL <strong>de</strong><br />

solution Tri-Reagent et broyés à l’ai<strong>de</strong> d’un broyeur à billes. A l’homogénat sont ajoutés<br />

0,2 mL <strong>de</strong> chloroforme <strong>pour</strong> provoquer une émulsion afin <strong>de</strong> séparer la fraction ARN <strong>de</strong>s<br />

fractions ADN/protéines/lipi<strong>de</strong>s. Après centrifugation (12 000 g, 15 min), les ARN totaux<br />

du surnageant sont précipités par ajout <strong>de</strong> 0,5 mL d’isopropanol. Les ARN totaux sont<br />

ensuite culottés par centrifugation (12 000 g, 10 min), lavés dans <strong>de</strong> l’alcool 70°, séchés et<br />

repris dans 50 microlitres d’eau-DEPC (incubation 15 min à 55°C).<br />

Dans le cadre <strong>de</strong>s banques Bathymodiolus et Lepetodrilus, les ARN messagers<br />

(ARNm) sont purifiés sur une résine oligodT à l’ai<strong>de</strong> du kit Promega PolyATtract mRNA<br />

Isolation Systems. Les ARN totaux sont hybridés à l’ai<strong>de</strong> d’une son<strong>de</strong> Oligo(dT) biotinylé<br />

(50 nmol/mL) en présence d’une solution SSC 0,5X. Les ARNm hybridés sont ensuite<br />

capturés sur <strong>de</strong>s sphères magnétiques <strong>de</strong> streptavidine couplées à un anticorps <strong>de</strong> la biotine<br />

et isolés du « pool » <strong>de</strong>s ARNs par aimantation. Les ARNm-sphères sont ensuite lavés<br />

plusieurs fois dans <strong>de</strong>s solutions 0,5X SSC et 0,1X SSC et resuspendus dans 1 mL d’eau-<br />

DEPC. Les ADNc sont synthétisés par rétrotranscription (8 µl <strong>de</strong> tampon contenant 1,2 µl<br />

<strong>de</strong> MgCl2 25mM, 4 µl <strong>de</strong> tampon <strong>de</strong> transcriptase reverse, 3 µl <strong>de</strong> dNTP 2mM, 100 U <strong>de</strong><br />

transcriptase reverse, une heure à 37°C) puis amplifiés, clonés et séquencés.<br />

3. Acquisition <strong>de</strong>s séquences<br />

3.1. Séquences du gène mitochondrial Cytochrome Oxydase I<br />

L’ADN mitochondrial étant généralement à hérédité uniquement maternelle (mais<br />

voir Skibinsky et al 1994 <strong>pour</strong> les exceptions), il est théoriquement présent en une seule<br />

copie (pas <strong>de</strong> recombinaison) au sein d’un individu. C’est <strong>pour</strong>quoi le séquençage du gène<br />

mtCOI a été effectué en direct, c’est-à-dire sans clonage préalable, à la plateforme<br />

Génomer <strong>de</strong> la <strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>. Le fragment du gène COI est amplifié par<br />

PCR à l’ai<strong>de</strong> d’amorces universelles (Folmer et al 1994) ou redéfinies à partir <strong>de</strong> ces<br />

amorces, puis purifié sur membrane Millipore et séquencé sur ABI.<br />

47


Chapitre 2<br />

3.2. Séquences <strong>de</strong>s gènes nucléaires : Marquage-Recapture (MR) <strong>de</strong><br />

séquences<br />

Les locus génomiques étudiés sont diploï<strong>de</strong>s. Les individus séquencés peuvent donc<br />

être homozygotes ou hétérozygotes au locus d’intérêt. Afin <strong>de</strong> ne pas brouiller le signal <strong>de</strong><br />

la séquence (présence d’in<strong>de</strong>ls ou nombreuses substitutions entre les allèles), il est<br />

nécessaire, après amplification par PCR et purification du produit, d’isoler les différents<br />

allèles par clonage. Le clonage étant un protocole relativement long et coûteux, il n’est pas<br />

possible d’effectuer <strong>de</strong>s clonages individuels <strong>pour</strong> chaque locus. Nous avons donc choisi<br />

une solution alternative : combiner les produits d’amplification <strong>de</strong>s individus appartenant à<br />

une même espèce, voire une même population, après les avoir préalablement marqués<br />

individuellement, et cloner l’ensemble. Pour éviter <strong>de</strong> biaiser la diversité allélique par<br />

clonage, les allèles sont donc « taggés » selon leur appartenance à un individu (Fig. 2.3)<br />

technique <strong>de</strong> marquage / re-capture <strong>de</strong> séquences : Bierne et al 2007, article présenté en<br />

annexe 1). Le marquage <strong>de</strong> la séquence consiste à ajouter quelques nucléoti<strong>de</strong>s permettant<br />

une i<strong>de</strong>ntification postérieure à l’extrémité 5’ <strong>de</strong>s amorces utilisées <strong>pour</strong> l’amplification<br />

<strong>de</strong>s individus. Ces nucléoti<strong>de</strong>s forment ainsi un « tag » qui, bien que ne s’hybridant pas<br />

avec le fragment d’ADN recherché, est amplifié par la Taq polymérase lors <strong>de</strong> la PCR.<br />

Ainsi, les allèles d’un individu amplifié sont i<strong>de</strong>ntifiés par une combinaison <strong>de</strong> « tags »,<br />

lesquels diffèrent d’un individu à l’autre. Ces « tags » sont ensuite retrouvés sur les<br />

séquences clonées <strong>de</strong> part et d’autre <strong>de</strong> l’insert, ce qui permet, a posteriori, d’attribuer les<br />

séquences aux individus (re-capture).<br />

48


Chapitre 2<br />

Brin d’ADN<br />

ACAG<br />

ACAG<br />

Amplification<br />

par PCR<br />

ACAG<br />

AG<br />

ACAG<br />

Chaque individu<br />

est amplifié avec<br />

un couple <strong>de</strong> tags<br />

différents<br />

Produits <strong>de</strong> PCR<br />

mixés dans un<br />

même tube<br />

1 clonage<br />

unique<br />

Séquençage<br />

avec un couple<br />

d’amorces<br />

plasmidiques<br />

Séquences<br />

réattribuées aux<br />

individus grâce à<br />

leurs tags<br />

1 couleur =1 puit = 1 individu<br />

= 1 couple <strong>de</strong> tags<br />

Séquence insérée dans<br />

un plasmi<strong>de</strong><br />

Fig 2.3 Schéma du fonctionnement <strong>de</strong> la métho<strong>de</strong> <strong>de</strong> marquage-recapture <strong>de</strong><br />

séquences utilisée <strong>pour</strong> le séquençage <strong>de</strong>s gènes nucléaires.<br />

49


Chapitre 2<br />

Des mutations artéfactuelles sur les « tags » ou <strong>de</strong> la recombinaison in vitro pouvant<br />

engendrer une mauvaise réassignation <strong>de</strong>s séquences, <strong>de</strong>s précautions particulières ont été<br />

prises afin <strong>de</strong> :<br />

(1) Détecter les mutations artéfactuelles sur les tags. Le choix <strong>de</strong>s tags a été réalisé<br />

afin que chaque tag (<strong>de</strong> 4 nucléoti<strong>de</strong>s) diffère d’au moins 2 substitutions permettant ainsi<br />

<strong>de</strong> repérer une éventuelle mutation artéfactuelle sur l’un <strong>de</strong>s tags.<br />

(2) Détecter les recombinaisons in vitro. Chaque individu a été amplifié avec un<br />

couple d’amorces présentant <strong>de</strong>s « tag » différents (individu 1 : F1R1, individu 2 : F2R2,<br />

Fig. 2.3). En conséquence, une séquence recapturée présentant <strong>de</strong>s combinaisons <strong>de</strong> tags<br />

non planifiées (F1R2 par exemple) est considérée comme recombinée in vitro.<br />

Pour chaque espèce et chaque gène, nous avons choisi <strong>de</strong> séquencer 16 individus par<br />

champ hydrothermal en effectuant 2 clonages par marquage recapture <strong>de</strong>s séquences issues<br />

<strong>de</strong> 8 individus chacun, permettant ainsi d’augmenter le nombre <strong>de</strong> séquences recapturées<br />

(en comparaison à 1 seul clonage <strong>de</strong> 16 individus). Trente <strong>de</strong>ux clones par clonage ont<br />

ensuite été séquencés au génoscope d’Evry.<br />

4. Traitement <strong>de</strong>s séquences<br />

4.1. Alignement et correction <strong>de</strong>s séquences liées aux erreurs <strong>de</strong> lecture<br />

du séquenceur<br />

Une fois le séquençage effectué par les plateformes <strong>de</strong> Genomer et du Génoscope,<br />

les régions plasmidiques (<strong>pour</strong> les gènes nucléaires) sont éliminées manuellement. Pour<br />

chaque gène, les séquences sens (forward) et antisens (reverse) sont alignées via ClustalW<br />

(Thompson et al 1994) sous Bioedit version 6.0.6 (Hall 1999) et l’alignement est vérifié<br />

visuellement. Les chromatogrammes <strong>de</strong> chaque séquence sont également alignés entre eux<br />

via la version <strong>de</strong> démonstration <strong>de</strong> CodonCo<strong>de</strong> aligner 2.0.6<br />

(http://www.codonco<strong>de</strong>.com/aligner/). L’alignement <strong>de</strong> ces chromatogrammes permet <strong>de</strong><br />

détecter une zone <strong>de</strong> recouvrement commune à l’ensemble <strong>de</strong>s séquences <strong>pour</strong> laquelle les<br />

50


Chapitre 2<br />

chromatogrammes concor<strong>de</strong>nt parfaitement entre chaque paire <strong>de</strong> séquences sens et<br />

antisens. La correspondance entre les bases inscrites par le séquenceur et les pics <strong>de</strong>s<br />

chromatogrammes est alors vérifiée jusqu’à cette zone <strong>de</strong> recouvrement incluse (sens : <strong>de</strong><br />

la première base jusqu’à la fin <strong>de</strong> la zone <strong>de</strong> recouvrement, antisens : début <strong>de</strong> la zone <strong>de</strong><br />

recouvrement jusqu’à la fin <strong>de</strong> la séquence). Dr Matthieu Bruneaux et moi-même avons<br />

développé un script sous C permettant <strong>de</strong> « rabouter » les séquences sens et antisens <strong>de</strong><br />

manière automatique en fournissant un fichier <strong>de</strong> séquences sens, un fichier antisens<br />

(format fasta) et la position du site <strong>de</strong> coupure (i<strong>de</strong>ntique <strong>pour</strong> toutes les séquences) défini<br />

par l’utilisateur dans la zone <strong>de</strong> recouvrement (cf annexe 2). Ce programme n’effectue pas<br />

<strong>de</strong> consensus <strong>de</strong> séquences, ne modifiant pas les séquences fournies. Plusieurs centaines <strong>de</strong><br />

séquences d’un même gène peuvent ainsi être raboutées en quelques secon<strong>de</strong>s. Ce<br />

programme a donc été utilisé indépendamment <strong>pour</strong> chaque gène et chaque espèce. Pour<br />

les séquences nucléaires, effectuées selon la métho<strong>de</strong> <strong>de</strong> marquage-recapture <strong>de</strong> séquences,<br />

nous avons ensuite réattribué chaque séquence à son individu amplifié par l’observation du<br />

couple <strong>de</strong> « tags » (Bierne et al 2007). Pour chaque gène, les bases correspondant aux<br />

amorces sont ensuite supprimées <strong>pour</strong> les analyses afin <strong>de</strong> ne pas biaiser les estimations <strong>de</strong><br />

polymorphisme et <strong>de</strong> divergence (zone conservée du fait <strong>de</strong> l’amorce).<br />

4.2. Recombinaison in vitro détectée via la métho<strong>de</strong> MR<br />

En utilisant un couple <strong>de</strong> tags différents <strong>pour</strong> chaque individu, nous avons pu mettre<br />

en évi<strong>de</strong>nce 22 séquences présentant une combinaison anormale <strong>de</strong> tags, indiquant que ces<br />

séquences sont issues d’un événement <strong>de</strong> recombinaison in vitro entre individus différents<br />

d’un même clonage. Le <strong>pour</strong>centage <strong>de</strong> séquences détectées recombinées in vitro entre<br />

individus est compris entre 0 et 1,82%, cette <strong>de</strong>rnière valeur ayant été obtenue <strong>pour</strong> la<br />

Globine X d’A. pompejana (Tableau 2.1).<br />

51


Chapitre 2<br />

Tableau 2.1 Pourcentage <strong>de</strong> séquences détectées recombinées in vitro entre différents<br />

individus.<br />

SAHH, S-Adénosyl Homocystéine Hydrolase ; PGM, Phosphoglucomutase ; VDG3,<br />

Veliger Digestive Gland 3.<br />

Espèce<br />

A. pompejana<br />

B. thermophilus<br />

L. elevatus<br />

Gène<br />

Taille moyenne<br />

du gène (pb)<br />

Pourcentage <strong>de</strong> séquences détectées<br />

recombinées in vitro entre individus<br />

SAHH 579 0,18<br />

Globine X 668 1,82<br />

PGM 1004 0,58<br />

SAHH 1005 0<br />

Lysozyme 1302 0,29<br />

Sulfotransférase 275 0<br />

Lysozyme 782 0,33<br />

Textilinine 1227 1,64<br />

VDG3 682 0,18<br />

Aucune corrélation n’a pu être détectée entre l’espèce ou la taille du gène séquencé<br />

et le <strong>pour</strong>centage <strong>de</strong> recombinaison in vitro détectée. Ces séquences recombinées ont été<br />

supprimées <strong>de</strong>s différents jeux <strong>de</strong> donnée.<br />

4.3. Correction du jeu <strong>de</strong> données : suppression <strong>de</strong>s singletons<br />

artéfactuels engendrant un multiallélisme<br />

Des erreurs d’amplification peuvent engendrer la présence <strong>de</strong> singletons<br />

artéfactuels dans les séquences (Cronn et al 2002) soit par erreur <strong>de</strong> réplication <strong>de</strong> la Taq<br />

polymérase ou lors <strong>de</strong> recombinaison entre allèles si l’ADN est <strong>de</strong> mauvaise qualité. À cela<br />

s’ajoute les mutations somatiques accumulées au cours <strong>de</strong> la croissance <strong>de</strong> l’organisme et<br />

non-transmises dans les lignées germinales. Si ces erreurs d’amplification ne sont<br />

généralement pas visibles lors d’un séquençage direct <strong>de</strong> produit <strong>de</strong> PCR (nombre <strong>de</strong><br />

copies important, masquant ces erreurs), le passage par une étape <strong>de</strong> clonage, en isolant les<br />

copies, les met en évi<strong>de</strong>nce. Ainsi, <strong>pour</strong> un gène <strong>de</strong> 1000 pb, les séquences diffèrent<br />

généralement entre elles d’environ un à <strong>de</strong>ux singletons artéfactuels. C’est <strong>pour</strong>quoi nous<br />

avons choisi d’appliquer une métho<strong>de</strong> <strong>de</strong> correction systématique <strong>de</strong>s séquences<br />

nucléaires. Dans un premier temps, trois fichiers <strong>de</strong> « qualité » différente sont réalisés :<br />

52


Chapitre 2<br />

(1) jeu qualité 1 (Fig. 2.4). Contient l’ensemble <strong>de</strong>s séquences recapturées, corrigées et<br />

réattribuées à un individu.<br />

(2) jeu qualité 2 (Fig. 2.4). Pour chaque individu, une séquence consensus <strong>de</strong>s<br />

séquences peu différentes entre elles (quelques singletons) est réalisée. Il est noté entre<br />

parenthèses le nombre <strong>de</strong> séquences ayant permis <strong>de</strong> réaliser ce consensus <strong>de</strong> la manière<br />

suivante : premier chiffre correspondant au nombre <strong>de</strong> séquences strictement i<strong>de</strong>ntiques à<br />

la séquence consensus, puis les autres chiffres correspon<strong>de</strong>nt aux autres groupes <strong>de</strong><br />

séquences différentes. Par exemple, cons (3 1 2) signifie : 3 séquences recapturées<br />

strictement i<strong>de</strong>ntiques à la séquence consensus, puis 1 séquence différente recapturée une<br />

seule fois, puis 1 autre séquence différente <strong>de</strong>s précé<strong>de</strong>ntes recapturée 2 fois. De même,<br />

cons (0 1 1 1) signifie qu’aucune séquence ne correspond strictement à la séquence<br />

consensus, le consensus a été fait sur la base <strong>de</strong> 3 séquences proches recapturées une seule<br />

fois chacune. Théoriquement (i.e. en absence <strong>de</strong> co-amplification <strong>de</strong> gènes), les fichiers <strong>de</strong><br />

qualité 2 comptent donc ainsi au maximum <strong>de</strong>ux séquences par individu recapturé<br />

correspondant aux 2 allèles <strong>de</strong> cet individu et sont utilisés lors <strong>de</strong> la réalisation <strong>de</strong>s arbres<br />

et réseaux phylogénétiques (permettant <strong>de</strong> visualiser ainsi <strong>de</strong> potentiels individus<br />

hybri<strong>de</strong>s).<br />

(3) jeu qualité 3 (Fig. 2.4). Une seule séquence par individu est sélectionnée : la<br />

séquence consensus issue du plus grand nombre <strong>de</strong> recaptures. Ce jeu est utilisé <strong>pour</strong><br />

effectuer les tests <strong>de</strong> neutralité <strong>pour</strong> éviter <strong>de</strong> biaiser la diversité nucléotidique observée, le<br />

nombre d’allèles recapturés n’étant pas systématiquement égal à 2 <strong>pour</strong> tous les individus.<br />

53


Chapitre 2<br />

Qualité 1 Qualité 2<br />

0,0005<br />

Qualité 3<br />

0,0005<br />

Qualité 4<br />

0,001<br />

0,0005<br />

Fig. 2.4 Arbres phylogénétiques réalisés à partir <strong>de</strong> six individus sur le gène<br />

Lysozyme <strong>de</strong> B. thermophilus illustrant les différents jeux <strong>de</strong> qualité obtenus<br />

Chaque couleur représente un individu différent.<br />

5. Traitement <strong>de</strong> la recombinaison<br />

5.1. Biais liés à la recombinaison dans les analyses <strong>de</strong> coalescence<br />

Outre les recombinants in vitro (artéfactuels), la recombinaison peut avoir une réalité<br />

biologique. Cependant, la plupart <strong>de</strong>s analyses basées sur la théorie <strong>de</strong> la coalescence pose<br />

l’hypothèse <strong>de</strong> l’absence <strong>de</strong> recombinaison. En effet, la recombinaison augmente la<br />

diversité génétique observée et modifie la forme du coalescent en établissant <strong>de</strong>s nœuds<br />

intermédiaires entre 2 allèles <strong>pour</strong> lesquels une divergence existe (rajeunissement <strong>de</strong> la<br />

divergence : Shriner et al 2003). Les sites recombinés présentent une diversité plus élevée<br />

que sous l’attendu neutre sans recombinaison, pouvant ainsi biaiser les analyses utilisant<br />

54


Chapitre 2<br />

<strong>de</strong>s indices <strong>de</strong> diversité génétique. Avant d’effectuer les analyses <strong>de</strong> coalescence, il est<br />

donc nécessaire <strong>de</strong> détecter et <strong>de</strong> supprimer du jeu <strong>de</strong> données les séquences<br />

potentiellement recombinantes.<br />

5.2. Détection et suppression <strong>de</strong>s séquences recombinantes<br />

Afin <strong>de</strong> supprimer les séquences présentant <strong>de</strong> la recombinaison, nous avons tout<br />

d’abord utilisé le « four-gamete test » <strong>de</strong> Hudson & Kaplan (1985) via le logiciel DNAsp<br />

4.10.3 (Rozas et al 2003) afin <strong>de</strong> détecter le nombre minimum <strong>de</strong> points <strong>de</strong> recombinaison<br />

(R m ) au sein du jeu <strong>de</strong> séquences et les sites nucléotidiques incriminés dans ces<br />

évènements <strong>de</strong> recombinaison. Une mutation reverse (homoplasique) peut également être<br />

comptée comme un événement <strong>de</strong> recombinaison par ce test (un point unique <strong>de</strong> mutation<br />

pouvant être compté comme événement <strong>de</strong> recombinaison s’il n’est pas en équilibre <strong>de</strong><br />

liaison avec les sites adjacents comparés, observation personnelle). Ces sites <strong>de</strong><br />

recombinaison sont comparés <strong>de</strong>ux à <strong>de</strong>ux <strong>pour</strong> l’ensemble <strong>de</strong>s séquences et les séquences<br />

détectées recombinantes (ou présentant <strong>de</strong>s sites homoplasiques) sont éliminées du jeu <strong>de</strong><br />

données (appelé <strong>de</strong> qualité 4, Fig. 2.4) jusqu’à atteindre R m = 0. Ce jeu <strong>de</strong> qualité 4 est<br />

alors utilisé <strong>pour</strong> les analyses <strong>de</strong> polymorphisme et <strong>de</strong> divergence posant l’hypothèse <strong>de</strong><br />

l’absence <strong>de</strong> recombinaison.<br />

6. Détection <strong>de</strong> gènes co-amplifiés (potentiellement paralogues)<br />

6.1. Bathymodiolus thermophilus<br />

6.1.1. Lysozyme<br />

Faure (2008) a mis en évi<strong>de</strong>nce la présence <strong>de</strong> <strong>de</strong>ux gènes paralogues sur le<br />

lysozyme, ces gènes étant co-amplifiés avec les amorces utilisées (divergence <strong>de</strong> 4% entre<br />

ces paralogues). Cependant, cet auteur n’ayant recapturé qu’une séquence du <strong>de</strong>uxième<br />

paralogue sur l’EPR, nous n’avons pas redéfini d’amorces spécifiques du premier<br />

paralogue. Hormis quatre séquences supplémentaires correspondant au <strong>de</strong>uxième<br />

paralogue, toutes les séquences obtenues sur le Lysozyme correspondaient au paralogue 1.<br />

Les analyses ont donc été effectuées uniquement sur le paralogue 1 dans le chapitre 3.<br />

55


Chapitre 2<br />

6.1.2. Sulfotransférase<br />

Certains individus recapturés sur le gène <strong>de</strong> la Sulfotransférase présentaient trois à<br />

quatre allèles distincts (multiples substitutions) par individu, laissant penser que nous<br />

avions co-amplifié au moins <strong>de</strong>ux gènes avec le couple d’amorces utilisé. Deux cla<strong>de</strong>s<br />

divergents <strong>de</strong> 2,6% ont été i<strong>de</strong>ntifiés (Fig. 2.5). La répartition <strong>de</strong>s séquences entre<br />

individus présentant plus <strong>de</strong> <strong>de</strong>ux allèles distincts nous a permis <strong>de</strong> poser l’hypothèse<br />

d’une paralogie associée à ces <strong>de</strong>ux cla<strong>de</strong>s. En effet, certains individus comptaient <strong>de</strong>ux<br />

allèles dans le cla<strong>de</strong> 1 et <strong>de</strong>ux autres dans le cla<strong>de</strong> 2, mais aucun individu ne possédait plus<br />

<strong>de</strong> <strong>de</strong>ux allèles recapturés au sein d’un même cla<strong>de</strong>. De plus, le cla<strong>de</strong> 1 présentait une<br />

divergence génétique entre les individus échantillonnés au Nord et au Sud <strong>de</strong> la barrière au<br />

flux <strong>de</strong> gènes préalablement détectée sur d’autres gènes (Fig. 2.5). De même, les séquences<br />

du cla<strong>de</strong> 2 se regroupent préférentiellement selon leur géographie (en termes <strong>de</strong><br />

fréquences). Les cla<strong>de</strong>s 1 et 2 ont donc été considérés comme <strong>de</strong>ux gènes distincts,<br />

respectivement nommés Sulfotransférase 1 et 2, <strong>pour</strong> les analyses sur B. thermophilus du<br />

chapitre 3.<br />

56


Chapitre 2<br />

Cla<strong>de</strong> 1 : Sulfotransférase 1<br />

Cla<strong>de</strong> 2 : Sulfotransférase 2<br />

0,005<br />

Fig. 2.5 Arbre phylogénétique permettant l’i<strong>de</strong>ntification <strong>de</strong> <strong>de</strong>ux gènes co-amplifiés<br />

sur la Sulfotransférase.<br />

En bleu, individus échantillonnés à 9°50’N et 7°25’S.<br />

En rouge, individus échantillonnés <strong>de</strong> 14°S à 21°33’S.<br />

6.2. Lepetodrilus elevatus<br />

Matabos et al (2008b) a décrit l’existence <strong>de</strong> <strong>de</strong>ux espèces cryptiques <strong>de</strong> L. elevatus<br />

divergentes <strong>de</strong> 6,5% sur le gène mtCOI. Cette étu<strong>de</strong> portait sur trois sites : 13°N, 9°50’N et<br />

17°25’S. Alors que les <strong>de</strong>ux espèces sont présentes en sympatrie à 9°50’N, seule une<br />

espèce était échantillonnée à 13°N et la secon<strong>de</strong> à 17°25’S. Lors <strong>de</strong> mon stage <strong>de</strong> Master 2<br />

(Plouviez 2006), nous avons séquencé trois gènes nucléaires <strong>pour</strong> <strong>de</strong>s individus présents à<br />

13°N et 17°25’S et donc diagnostiques <strong>de</strong>s <strong>de</strong>ux espèces cryptiques. Nous fûmes alors<br />

confrontés à un problème <strong>de</strong> multiallélisme au niveau <strong>de</strong> ces gènes nucléaires, laissant<br />

penser à la co-amplification <strong>de</strong> plusieurs gènes. Durant la thèse, nous avons donc tenté <strong>de</strong><br />

57


Chapitre 2<br />

résoudre ces problèmes <strong>de</strong> multiallélisme afin d’utiliser ces gènes dans le cadre d’une<br />

étu<strong>de</strong> phylogéographique.<br />

6.2.1. Lysozyme<br />

Dans un premier temps, une quinzaine <strong>de</strong> L. elevatus ont été clonés et séquencés<br />

individuellement (pas d’utilisation <strong>de</strong> la métho<strong>de</strong> <strong>de</strong> MR). Pour chaque individu, nous<br />

avons séquencé une dizaine <strong>de</strong> clones afin <strong>de</strong> détecter les gènes co-amplifiés. Quatre<br />

cla<strong>de</strong>s ont ainsi pu être mis en évi<strong>de</strong>nce (Fig. 2.6).<br />

58<br />

Fig. 2.6 Arbre phylogénétique <strong>de</strong> type qualité 1 réalisé sur les séquences <strong>de</strong>s<br />

individus clonés individuellement sur le Lysozyme.<br />

En bleu et rouge, espèces cryptiques <strong>de</strong> type 13°N et 17°25’S (Matabos et al 2008),<br />

respectivement, sur la base du mtCOI sur la base du mtCOI (Matabos et al 2008). Les<br />

allèles présents sur la divergence constituent <strong>de</strong>s recombinants.


Chapitre 2<br />

Les séquences <strong>de</strong>s cla<strong>de</strong>s 1 et 2 présentent le même polymorphisme d’in<strong>de</strong>ls<br />

(insertions/délétions) et le même type exonique. Elles diffèrent essentiellement par <strong>de</strong>s<br />

substitutions situées dans l’intron. Le cla<strong>de</strong> 3 diffère notamment <strong>de</strong> ces <strong>de</strong>ux cla<strong>de</strong>s par un<br />

nombre <strong>de</strong> 13 substitutions non-synonymes fixées sur l’exon 2. Le cla<strong>de</strong> 4 possè<strong>de</strong> le<br />

même type d’exon que les cla<strong>de</strong>s 1 et 2. Il diffère <strong>de</strong>s autres cla<strong>de</strong>s par une délétion<br />

d’environ 560 pb au début <strong>de</strong> l’intron et une insertion <strong>de</strong> ∼260 pb à la fin <strong>de</strong> celui-ci, ces<br />

in<strong>de</strong>ls étant également en déséquilibre <strong>de</strong> liaison à <strong>de</strong>s substitutions introniques multiples.<br />

Certains individus comptaient <strong>de</strong>s séquences à la fois dans les cla<strong>de</strong>s 1, 3 et 4, indiquant la<br />

présence d’au moins <strong>de</strong>ux gènes co-amplifiés. Le cla<strong>de</strong> 2 étant essentiellement constitué <strong>de</strong><br />

séquences <strong>de</strong> l’espèce type 13°N, nous avons considéré que la divergence observée entre<br />

les cla<strong>de</strong>s 1 et 2 (Fig. 2.6) était due à la divergence entre espèces cryptiques et non entre<br />

gènes différents. Nous avons alors défini trois couples d’amorces discriminant les cla<strong>de</strong>s 1-<br />

2, 3 et 4 et testé ces amorces sur une cinquantaine d’individus par espèce. Pour les <strong>de</strong>ux<br />

espèces, nous avons ainsi mis en évi<strong>de</strong>nce <strong>de</strong>s individus amplifiant <strong>pour</strong> les trois couples<br />

d’amorces, confirmant l’existence d’au moins <strong>de</strong>ux gènes différents. Néanmoins,<br />

contrairement aux amorces définies <strong>pour</strong> les cla<strong>de</strong>s 1-2, les cla<strong>de</strong>s 3 et 4 n’amplifiaient pas<br />

sur l’ensemble <strong>de</strong>s individus, suggérant un mauvais échantillonnage <strong>de</strong> ces cla<strong>de</strong>s 3 et 4.<br />

Par conséquent, nous avons considéré que les cla<strong>de</strong>s 1 et 2 constituaient un gène unique et<br />

séquencé ce gène par la métho<strong>de</strong> <strong>de</strong> MR <strong>pour</strong> l’étu<strong>de</strong> phylogéographique. Les séquences<br />

obtenues à l’issue <strong>de</strong> ce séquençage se sont révélées compatibles avec un gène unique<br />

(<strong>de</strong>ux allèles maximum par individus échantillonnés).<br />

6.2.2. Veliger Digestive Gland 3 (VDG3)<br />

Le gène VDG3 présentait également une co-amplification d’allèles appartenant à<br />

plusieurs locus (allèles multiples) avec la présence d’au moins <strong>de</strong>ux cla<strong>de</strong>s divergents sans<br />

rapport avec la présence <strong>de</strong>s 2 espèces cryptiques. Nous avons donc redéfini une amorce<br />

sens afin <strong>de</strong> n’amplifier qu’un seul cla<strong>de</strong> (l’amorce antisens restant inchangée). Nous<br />

avons ensuite cloné individuellement les individus <strong>de</strong> chaque espèce, respectivement <strong>de</strong><br />

type 13°N et 17°25’S au locus mtCOI (Matabos et al 2008b) et séquencé 6-8 clones par<br />

individu. Les séquences obtenues comptaient un maximum <strong>de</strong> <strong>de</strong>ux allèles par individu,<br />

suggérant que les allèles observées proviennent d’un gène unique. Là encore les allèles se<br />

59


Chapitre 2<br />

regroupent selon les espèces cryptiques (Fig. 2.7). Nous avons donc utilisé ce couple<br />

d’amorces <strong>pour</strong> le séquençage par la métho<strong>de</strong> MR.<br />

0,005<br />

Fig. 2.7 Arbre phylogénétique <strong>de</strong> type qualité 1 réalisé sur les séquences <strong>de</strong>s<br />

individus clonés individuellement sur le VDG3 avec les amorces redéfinies.<br />

Chaque couleur représente un même individu. En bleu, séquences <strong>de</strong> <strong>de</strong>ux individus<br />

correspondant à l’espèce cryptique <strong>de</strong> type 13°N sur le mtCOI (Matabos et al 2008). En<br />

jaune, rouge et vert, les séquences d’individus <strong>de</strong> l’autre espèce cryptique. Chaque<br />

individu présente au maximum <strong>de</strong>ux allèles différents (séquences comptant un<br />

polymorphisme <strong>de</strong> singletons, e.g. individu en jaune, mais correspondant à un même allèle<br />

consensus <strong>pour</strong> le jeu <strong>de</strong> qualité 2)<br />

Les séquences obtenues par MR se distribuaient en trois cla<strong>de</strong>s majoritaires, avec<br />

<strong>de</strong>s séquences intermédiaires entre ces cla<strong>de</strong>s (Fig. 2.8.A). Si la plupart <strong>de</strong>s individus<br />

séquencés présentaient au maximum <strong>de</strong>ux allèles par individu, dix individus comptaient<br />

plusieurs allèles bien différents. Nous avons donc examiné, <strong>pour</strong> chacun <strong>de</strong> ces individus,<br />

s’il existait <strong>de</strong>s allèles recombinés. Pour ces dix individus, nous avons ainsi pu mettre en<br />

évi<strong>de</strong>nce la présence d’un ou <strong>de</strong>ux allèles appartenant à un <strong>de</strong>s 2 cla<strong>de</strong>s préalablement<br />

i<strong>de</strong>ntifiés ainsi que <strong>de</strong>s allèles recombinés intra-cla<strong>de</strong>, <strong>de</strong>s allèles recombinés entre les 2<br />

cla<strong>de</strong>s et, <strong>de</strong> façon encore plus inattendue, <strong>de</strong>s allèles recombinés entre le gène d’intérêt et<br />

le gène que nous avions choisi d’exclure par le biais <strong>de</strong> l’utilisation <strong>de</strong> la nouvelle amorce<br />

sens. Les ADN génomiques <strong>de</strong> nos individus étant assez dégradés, il est possible que la<br />

60


Chapitre 2<br />

Taq polymérase ait associé <strong>de</strong>s morceaux <strong>de</strong> séquences sens et antisens <strong>de</strong>s différents<br />

cla<strong>de</strong>s voire <strong>de</strong>s différents paralogues (l’amorce antisens n’ayant pas été redéfinie). Pour<br />

ces dix individus, après suppression <strong>de</strong>s séquences recombinées, nous avons obtenu un<br />

maximum <strong>de</strong> <strong>de</strong>ux allèles différents par individu et les cla<strong>de</strong>s <strong>de</strong> l’arbre phylogénétique se<br />

sont révélés mieux définis (Fig. 2.8.B). Nous avons donc considéré que ce gène constituait<br />

un gène unique et avons donc utilisé ces séquences <strong>pour</strong> l’analyse phylogéographique <strong>de</strong> L.<br />

elevatus dans le chapitre 4.<br />

A<br />

B<br />

Fig. 2.8 Arbre phylogénétique <strong>de</strong> type qualité 2 réalisé sur les séquences issues <strong>de</strong> la<br />

métho<strong>de</strong> MR sur le VDG3 comprenant les séquences recombinantes <strong>de</strong>s dix<br />

individus présentant plus <strong>de</strong> <strong>de</strong>ux séquences bien différentes (A), sans les séquences<br />

recombinantes <strong>de</strong> ces dix individus (B).<br />

61


Chapitre 2<br />

6.2.3. Textilinine<br />

Les séquences du gène Textilinine obtenues par la métho<strong>de</strong> MR se répartissent en<br />

quatre cla<strong>de</strong>s principaux (Fig. 2.9). Certains individus comptent plus <strong>de</strong> dix séquences , ce<br />

qui a permis la mise en évi<strong>de</strong>nce d’allèles multiples (plus <strong>de</strong> <strong>de</strong>ux allèles par individu),<br />

suggérant, là encore, la co-amplification <strong>de</strong> plusieurs gènes. Un individu recapturé douze<br />

fois possè<strong>de</strong> un allèle dans le cla<strong>de</strong> 1, <strong>de</strong>ux dans le cla<strong>de</strong> 2, et un dans le cla<strong>de</strong> 3, indiquant<br />

que le cla<strong>de</strong> 2 constituerait un gène différent <strong>de</strong>s <strong>de</strong>ux autres cla<strong>de</strong>s. De même, un autre<br />

individu recapturé treize fois présente un allèle dans le cla<strong>de</strong> 1, un dans le cla<strong>de</strong> 3 et <strong>de</strong>ux<br />

allèles dans le cla<strong>de</strong> 4, suggérant que le cla<strong>de</strong> 4 serait également différent <strong>de</strong>s cla<strong>de</strong>s 1 et 2.<br />

Nous n’avons pas recapturé d’individus rendant incompatible l’appartenance <strong>de</strong>s cla<strong>de</strong>s 1<br />

et 3 à un même gène. De plus, l’ensemble <strong>de</strong>s séquences recapturées provenant <strong>de</strong> l’espèce<br />

cryptique mitochondriale <strong>de</strong> type 13°N (Matabos et al 2008b) est distribué dans le cla<strong>de</strong> 1,<br />

indiquant que la divergence entre les cla<strong>de</strong>s 1 et 3 <strong>pour</strong>raient correspondre à la divergence<br />

attendue entre 2 espèces différentes. Les cla<strong>de</strong>s 2 et 4 <strong>pour</strong>raient constituer un <strong>de</strong>uxième<br />

gène co-amplifié, mais aucune séquence <strong>de</strong> l’espèce cryptique <strong>de</strong> type 13°N n’a été<br />

recapturée dans ces cla<strong>de</strong>s (amorces ne permettant pas d’amplifier cette espèce, ou<br />

apparition d’une duplication secondaire dans la lignée mitochondriale sud <strong>de</strong> L. elevatus).<br />

Par conséquent, bien que l’approche reste critiquable, nous avons considéré que les cla<strong>de</strong>s<br />

1 et 3 pouvaient constituer un gène unique utilisable <strong>pour</strong> une approche <strong>de</strong><br />

phylogéographie sur L. elevatus.<br />

62


Chapitre 2<br />

Fig. 2.9 Réseau réalisé en Median Joining <strong>de</strong>s séquences <strong>de</strong> Textilinine sur le jeu <strong>de</strong><br />

données <strong>de</strong> qualité 2 <strong>pour</strong> lequel <strong>de</strong>s individus présentent <strong>de</strong>s allèles multiples.<br />

Pour faciliter la représentation, le diamètre <strong>de</strong>s cercles n’est pas proportionnel au<br />

nombre <strong>de</strong> séquences recapturées. La longueur <strong>de</strong>s branches est proportionnelle à la<br />

divergence.<br />

7. Recherche d’une divergence partagée entre plusieurs espèces :<br />

MsBayes<br />

Lorsque plusieurs espèces présentent <strong>de</strong>ux cla<strong>de</strong>s divergents <strong>de</strong> part et d’autre d’une<br />

potentielle barrière aux flux <strong>de</strong> gène, il s’avère intéressant <strong>de</strong> tester si la divergence<br />

observée entre ces paires <strong>de</strong> cla<strong>de</strong>s est due à un événement d’isolement commun à toutes<br />

les espèces ou à plusieurs évènements différents et ainsi <strong>de</strong> dater ces évènements. Une<br />

nouvelle approche dite ‘Hierarchical Approximate Bayesian Computation’ (ABC) permet<br />

d’estimer, via le logiciel MsBayes (Hickerson et al 2006b), ce nombre d’évènements <strong>de</strong><br />

63


Chapitre 2<br />

divergence (hypervariable Ψ) et <strong>de</strong> les dater (temps <strong>de</strong> divergence moyen E(τ) et Ω =<br />

E(τ)/Var(τ)). Cette approche est basée sur <strong>de</strong>s simulations <strong>de</strong> statistiques résumées (e.g.<br />

diversité génétique, θ W ) à partir d’un modèle d’isolement en 2 populations sœurs évoluant<br />

sous l’hypothèse d’une variation indépendante <strong>de</strong> leur taille efficace (Fig. 2.10).<br />

Présent<br />

Population<br />

ancestrale<br />

Passé<br />

Fig. 2.10 Représentation schématique d’un modèle où une population ancestrale se<br />

sépare en <strong>de</strong>ux populations filles 1 et 2 au temps τ. Tirée <strong>de</strong> Hickerson et al (2006a).<br />

N A , N 1 et N 2 , tailles <strong>de</strong>s populations ancestrales, fille 1 actuelle et fille 2 actuelle,<br />

respectivement.<br />

N bott1 et N bott2 , tailles <strong>de</strong>s populations 1 et 2 après la séparation <strong>de</strong>s populations <strong>pour</strong> les<br />

temps τ bott1 , τ bott2<br />

τ, temps d’isolement <strong>de</strong>s populations<br />

64


Chapitre 2<br />

Cette approche se divise en trois étapes :<br />

(1) Un fichier contenant les séquences observées <strong>de</strong>s <strong>de</strong>ux cla<strong>de</strong>s divergents est fourni<br />

<strong>pour</strong> chaque espèce. À partir <strong>de</strong> ces jeux <strong>de</strong> séquences, le logiciel MsBayes estime<br />

une série <strong>de</strong> statistiques résumées <strong>pour</strong> chaque cla<strong>de</strong>/lignée divergente (Tableau<br />

2.2).<br />

Tableau 2.2 Statistiques résumées estimés à partir <strong>de</strong>s séquences observées ou<br />

simulées. Tiré <strong>de</strong> Hickerson et al (2006a).<br />

Notation Statistique résumée<br />

Référence<br />

π Moyenne <strong>de</strong>s différences par paires <strong>de</strong> séquences Tajima 1983<br />

π b Moyenne <strong>de</strong>s différences par paires <strong>de</strong> séquences intercla<strong>de</strong>s Takahata & Nei<br />

1985<br />

π w Moyenne <strong>de</strong>s différences par paires <strong>de</strong> séquences intracla<strong>de</strong>s Takahata & Nei<br />

1985<br />

π net Moyenne nette <strong>de</strong>s différences par paires <strong>de</strong> séquences entre<br />

cla<strong>de</strong>s (=π w - π b )<br />

Takahata & Nei<br />

1985<br />

S Nombre <strong>de</strong> sites polymorphes Watterson 1975<br />

θ w Nombre <strong>de</strong> sites polymorphes normalisés par la taille <strong>de</strong> Watterson 1975<br />

l’échantillonnage (thêta <strong>de</strong> Watterson)<br />

D D <strong>de</strong> Tajima Tajima 1989<br />

(2) 100 000 hyper-paramètres (inter-espèces : Ψ, E(τ) et Ω) et sous-paramètres (intraespèce<br />

: N A , N 1 , N 2 , N bott1 , N bott2 , τ bott1 , τ bott2 , τ, calculés <strong>pour</strong> chaque espèce selon le<br />

modèle présenté Fig. 2.10) sont aléatoirement tirés du modèle d’isolement avec goulôts<br />

d’étranglements dans les lignées produites. Chacune <strong>de</strong> ces séries <strong>de</strong> paramètres permet<br />

<strong>de</strong> simuler un jeu différent <strong>de</strong> données <strong>de</strong> séquences (100 000 jeux <strong>de</strong> séquences) <strong>pour</strong><br />

chaque espèce (paires <strong>de</strong> cla<strong>de</strong>s). Les statistiques résumées sont ensuite calculées <strong>pour</strong><br />

chaque jeu <strong>de</strong> séquences simulé<br />

(3) Les statistiques résumées calculées à partir <strong>de</strong>s jeux <strong>de</strong> séquences simulés sont<br />

ensuite comparées à celles <strong>de</strong>s séquences observées. Les distributions <strong>de</strong>s probabilités<br />

postérieures <strong>de</strong>s hyper-paramètres (Ψ, E(τ) et Ω) et sous-paramètres (N A , N 1 , N 2 , N bott1 ,<br />

N bott2 , τ bott1 , τ bott2 , τ) associés proviennent <strong>de</strong>s 1000 statistiques résumées qui<br />

s’approximent le mieux aux statistiques résumées observées.<br />

65


Chapitre 2<br />

8. Migration à travers une barrière aux flux <strong>de</strong> gènes : Migrate-n<br />

!<br />

!<br />

La migration entre populations peut être analysée via <strong>de</strong>s logiciels tels que Migrate-n<br />

(Beerli & Felsenstein 2001), et notamment permettre l’analyse <strong>de</strong> cette migration à travers<br />

une barrière géographique. Migrate-n utilise une approche basée sur la théorie <strong>de</strong> la<br />

coalescence afin d’estimer les valeurs les plus vraisemblables <strong>de</strong>s taux <strong>de</strong> migration entre<br />

populations et celles <strong>de</strong> leur taille efficace (Beerli & Felsenstein 1999, 2001). La matrice<br />

$ " 1<br />

M 2,1<br />

M 3,1<br />

... M n,1<br />

'<br />

&<br />

)<br />

M 1,2<br />

" 2<br />

M 3,2<br />

... M n,2<br />

P = &<br />

) contient l’ensemble <strong>de</strong> ces paramètres estimés<br />

& ... ... ... ... ... )<br />

&<br />

)<br />

% M 1,n<br />

M 2,n<br />

... M n#1,n<br />

" n (<br />

(Fig. 2.11), avec M i,j correspondant au taux <strong>de</strong> migration <strong>de</strong> la population i vers j divisé par<br />

le taux <strong>de</strong> mutations par génération (µ), et Θ i correspondant au paramètre démographique<br />

<strong>de</strong> la population i (= 4 Ne i µ avec Ne i la taille efficace <strong>de</strong> la population i). Les valeurs <strong>de</strong><br />

vraisemblance <strong>de</strong> cette matrice (L(P)) sont calculées selon la relation<br />

L(P) = " prob(G /P).prob(D/G) avec prob(G/P), la probabilité d’observer la généalogie<br />

G<br />

<strong>pour</strong> les paramètres P choisis et prob(D/G) la probabilité d’obtenir le jeu <strong>de</strong> données<br />

observé <strong>pour</strong> une généalogie simulée.<br />

M 1,2<br />

Θ 1 Θ 2<br />

M 2,1<br />

M 3,1<br />

M 1,3<br />

M 2,3<br />

M 3,2<br />

Θ 3<br />

Fig. 2.11 Représentation graphique du modèle utilisée par Migrate-n.<br />

Exemple <strong>de</strong> trois populations (cercles), les valeurs <strong>de</strong> M i,j pouvant être nulles<br />

66


Chapitre 2<br />

Une approche Metropolis-Hastings <strong>de</strong> chaînes <strong>de</strong> Markov Chain selon une procédure<br />

Monte Carlo (MCMC, Metropolis et al 1953) est utilisée afin d’échantillonner les valeurs<br />

<strong>de</strong> P les plus probables dans l’espace <strong>de</strong>s paramètres possibles (i.e. espace <strong>de</strong>s coalescents<br />

possibles). Cette approche MCMC consiste à estimer la vraisemblance <strong>de</strong> paramètres L(P)<br />

aléatoirement choisis dans cet espace. Les paramètres <strong>de</strong> cette matrice <strong>de</strong> départ sont<br />

ensuite légèrement modifiés <strong>pour</strong> donner une nouvelle matrice <strong>de</strong> paramètres et sa valeur<br />

<strong>de</strong> vraisemblance associée est comparée à la précé<strong>de</strong>nte :<br />

(1) Si L(P) nouvelle matrice > L(P) matrice <strong>de</strong> départ alors le processus d’exploration <strong>de</strong>s<br />

paramètres se <strong>pour</strong>suit en estimant une autre matrice.<br />

(2) Lorsque L(P) matrice n < L(P) matrice n-1 , alors le processus <strong>de</strong> recherche <strong>de</strong>s<br />

paramètres P est stoppé et les paramètres <strong>de</strong> la matrice n-1 sont considérés<br />

comme les plus vraisemblables.<br />

Afin <strong>de</strong> mieux explorer l’univers <strong>de</strong>s possibles, ce processus est effectué sur plusieurs<br />

matrices <strong>de</strong> départ, permettant ainsi <strong>de</strong> s’assurer que le choix <strong>de</strong> la matrice la plus<br />

vraisemblable converge entre les simulations et ainsi <strong>de</strong> maximiser la probabilité que cette<br />

matrice reflète les patrons <strong>de</strong> migration réels entre populations.<br />

En estimant les valeurs les plus probables <strong>de</strong>s taux <strong>de</strong> migration souvent<br />

assymétriques entre populations, le logiciel Migrate-n peut ainsi permettre <strong>de</strong> mieux<br />

appréhen<strong>de</strong>r le sens <strong>de</strong> la migration entre populations et notamment au travers une barrière<br />

aux flux <strong>de</strong> gènes. Néanmoins, ce programme fait l’hypothèse <strong>de</strong> l’équilibre<br />

mutation/dérive (pas <strong>de</strong> changements démographiques au cours du temps) et d’absence <strong>de</strong><br />

recombinaison intralocus (et interlocus lorsque plusieurs gènes sont utilisés dans une<br />

même simulation : série <strong>de</strong> génotypes <strong>pour</strong> un même individu).<br />

Le logiciel Migrate-n permet donc <strong>de</strong> comparer les compositions génétiques <strong>de</strong>s<br />

régions géographiques et peut permettre <strong>de</strong> définir <strong>de</strong>s taux <strong>de</strong> migration entre ces régions.<br />

Cette approche ne fait pas d’hypothèse sur la spéciation en cours. Ainsi, lorsque <strong>de</strong>ux<br />

entités génétiques différentes sont présentes au sein d’une même population, les individus<br />

migrants mais aussi les individus potentiellement introgressés entre ces entités seront<br />

comptabilisés dans le calcul du taux <strong>de</strong> migration.<br />

67


Chapitre 2<br />

9. Modèle d’ « Isolement avec Migration »<br />

La séparation d’une population ancestrale en <strong>de</strong>ux populations filles divergentes <strong>de</strong><br />

part et d’autre d’une barrière physique ou génétique est un processus long qui peut se<br />

réaliser en absence totale ou en présence d’un flux <strong>de</strong> gènes limité. Le modèle d’Isolement<br />

avec Migration (logiciel IM et IMa, Nielsen & Wakeley 2001, Hey & Nielsen 2004, 2007)<br />

permet <strong>de</strong> distinguer l’effet relatif <strong>de</strong> la divergence et <strong>de</strong>s flux <strong>de</strong> gènes dans le processus<br />

ayant conduit à l’isolement génétique <strong>de</strong> lignées. Une approche MCMC est utilisée afin<br />

d’estimer les probabilités postérieures <strong>de</strong>s différents paramètres du modèle (Fig. 2.12)<br />

soient:<br />

(1) θ A , θ 1 , et θ 2 , les paramètres démographiques <strong>de</strong>s populations ancestrale,<br />

filles 1 et 2, respectivement (avec θ i = 4N i u <strong>pour</strong> un gène nucléaire et 2 N i<br />

u <strong>pour</strong> un gène mitochondrial, N i correspondant à la taille <strong>de</strong> la population i<br />

considérée)<br />

(2) La proportion d’individus <strong>de</strong> la population 1 qui est remplacée à chaque<br />

génération par <strong>de</strong>s migrants <strong>de</strong> la population 2 (m1) et vice-versa (m2) et<br />

(3) Le temps <strong>de</strong> début <strong>de</strong> séparation <strong>de</strong>s lignées (t).<br />

Présent<br />

Population 1 Population 2<br />

Passé<br />

Population<br />

ancestrale<br />

Fig. 2.12 Représentation graphique du modèle d’Isolement avec Migration.<br />

Tirée <strong>de</strong> Nielsen & Wakeley (2001).<br />

u, taux <strong>de</strong> mutation par génération.<br />

68


Chapitre 2<br />

Ce modèle d’Isolement avec Migration peut ainsi permettre d’émettre <strong>de</strong>s<br />

hypothèses <strong>pour</strong> expliquer les flux <strong>de</strong> gènes potentiellement détectés. Ainsi, la présence <strong>de</strong><br />

flux <strong>de</strong> gènes entre les lignées <strong>pour</strong>rait correspondre à un isolement progressif <strong>de</strong>s lignées<br />

en présence <strong>de</strong> migration pendant le processus d’isolement (spéciation<br />

parapatrique/sympatrique) ou à un tri incomplet <strong>de</strong>s allèles (phase <strong>de</strong> spéciation toujours<br />

en cours). Au contraire, une absence <strong>de</strong> flux <strong>de</strong> gènes entre les lignées accrédite un<br />

processus <strong>de</strong> type allopatrique dans lequel la barrière est totalement imperméable. Il existe<br />

cependant un nombre <strong>de</strong> cas ou cette barrière cesse d’être imperméable et favorise la mise<br />

en place d’une zone <strong>de</strong> contact secondaire entre les lignées préalablement isolées. Dans ce<br />

cas, <strong>de</strong>s flux <strong>de</strong> gènes récents sont en général observés et traduisent <strong>de</strong>s phénomènes<br />

d’hybridation entre lignées avec une possibilité d’introgression d’allèles à certains locus si<br />

les croisements <strong>de</strong>s hybri<strong>de</strong>s avec les lignées parentales sont encore possibles (e.g. Faure et<br />

al 2009).<br />

La confrontation <strong>de</strong>s résultats <strong>de</strong> Migrate-n et du modèle d’Isolement avec<br />

Migration peut ainsi permettre d’évaluer la part <strong>de</strong> la migration liée à l’introgression<br />

(détectée par IM/IMa) <strong>de</strong> celle indépendante <strong>de</strong> la spéciation (uniquement prise en compte<br />

par Migrate-n).<br />

10. Écart à la neutralité : sélection versus démographie<br />

10.1. Approche multilocus visant à détecter un ou plusieurs locus sous<br />

sélection positive : test <strong>de</strong> Hudson Kreitman Aguadé<br />

Hudson, Kreitman et Aguadé (HKA, 1987) proposent un test statistique multilocus<br />

<strong>de</strong> polymorphisme/divergence basé sur la comparaison <strong>de</strong> séquences d’ADN intra- et interspécifiques.<br />

Un nombre L <strong>de</strong> locus sont échantillonnés respectivement chez <strong>de</strong>ux espèces<br />

proches. Soient S A i et S B i le nombre <strong>de</strong> sites nucléotidiques polymorphes respectivement<br />

chez les espèces A et B au locus i. et Di le nombre <strong>de</strong> différences nucléotidiques entre<br />

<strong>de</strong>ux séquences tirées au hasard chez chacune <strong>de</strong> ces <strong>de</strong>ux espèces au même locus i. Sous<br />

l’hypothèse neutraliste, S A i , S B<br />

i et D i ne sont soumis qu’à la dérive génétique. Si<br />

l’ensemble <strong>de</strong>s locus échantillonnés est neutre, alors la statistique<br />

69


Chapitre 2<br />

!<br />

L<br />

(S A<br />

X 2 = i<br />

" E^<br />

(S A i<br />

)) 2 L<br />

(S<br />

# B + i<br />

" E^<br />

(S B i<br />

)) 2 L<br />

(D<br />

^ # + i<br />

" E^<br />

(D i<br />

))<br />

i=1 Var(S A ^ # 2<br />

i<br />

) i=1 Var(S B ^<br />

suit une distribution <strong>de</strong> χ 2 à<br />

i<br />

) i=1 Var(D i<br />

)<br />

2L – 2 <strong>de</strong>grés <strong>de</strong> liberté. Au contraire, si cette statistique ne suit pas cette distribution, on<br />

peut en déduire qu’au moins un locus évolue sous sélection. Or, un locus soumis à la<br />

sélection darwinienne positive est supposé avoir un rapport polymorphisme sur divergence<br />

plus faible qu’un locus neutre. Par conséquent, le test HKA peut permettre <strong>de</strong> déterminer<br />

les locus assujettis à sélection darwinienne positive, sans <strong>pour</strong> autant donner la raison <strong>de</strong><br />

cette sélection. Néanmoins, ce test ne permet pas d’affirmer si le locus échantillonné le<br />

moins soumis à sélection est neutre ou simplement sous <strong>de</strong>s pressions <strong>de</strong> sélection plus<br />

faible. De plus, ce test fait plusieurs hypothèses pouvant être invali<strong>de</strong>s, en particulier, il<br />

suppose l’absence <strong>de</strong> recombinaison intralocus et <strong>de</strong> liaison inter-locus mais surtout <strong>de</strong>s<br />

tailles efficaces stationnaires <strong>de</strong>puis la spéciation <strong>de</strong>s <strong>de</strong>ux espèces (Hudson et al 1987).<br />

Or, cette <strong>de</strong>rnière est rarement vérifiée (Nei & Kumar 2000).<br />

10.2. Approche substitutions synonymes / non-synonymes en vu <strong>de</strong> la<br />

détection <strong>de</strong> sélection positive versus négative : test <strong>de</strong> MacDonald-<br />

Kreitman<br />

Le test <strong>de</strong> McDonald-Kreitman (MK, 1991a) repose également sur la comparaison<br />

intra- et inter-spécifique <strong>de</strong> séquences mais en comparant 2 classes <strong>de</strong> mutations : les<br />

mutations non-synonymes assujetties à la sélection et les mutations synonymes supposées<br />

neutres. Il s’agit d’un test utilisant une approche mono-locus. Il confronte le ratio <strong>de</strong><br />

polymorphisme p n / p s au ratio <strong>de</strong> divergence d n / d s (avec p n et p s le nombre <strong>de</strong> différences<br />

respectivement non-synonymes et synonymes entre séquences d’une même espèce, d n et d s<br />

le nombre <strong>de</strong> différences respectivement non synonymes et synonymes entre séquences<br />

d’espèces différentes mais proches). Si le locus étudié a évolué <strong>de</strong> façon neutre chez les<br />

espèces étudiées, alors ces rapports doivent être les mêmes: p n / p s = d n / d s . Si p n / p s < d n /<br />

d s , alors <strong>de</strong>s mutations non synonymes se sont fixées dans la divergence entre espèces sous<br />

l’effet <strong>de</strong> la sélection darwinienne positive. Si p n / p s > d n / d s , <strong>de</strong>s mutations non<br />

synonymes légèrement délétères ségrègent dans les populations mais ne fixent pas. Un<br />

simple test d’homogénéité permet <strong>de</strong> détecter l’écart à l’hypothèse <strong>de</strong> neutralité. Le test<br />

MK (1991a) permet <strong>de</strong> tester si un locus particulier est soumis à sélection sans nécessité<br />

l’accès à d’autres locus. De plus, ce test a <strong>pour</strong> avantage <strong>de</strong> pouvoir comparer plusieurs<br />

70


Chapitre 2<br />

espèces relativement proches et ne fait pas d’hypothèse sur la taille efficace <strong>de</strong>s<br />

populations. Cependant, Graur et Li (1991) et Whittam et Nei (1991) suggèrent qu’un<br />

faible nombre <strong>de</strong> séquences ou d’espèces échantillonnées peut sous estimer le ratio d n / d s .<br />

McDonald et Kreitman (1991b) rétorquent que cela n’affecterait pas la validité du test <strong>de</strong><br />

neutralité puisqu’un faible échantillonnage <strong>de</strong> séquences et d’espèces entraînerait une<br />

sous-estimation égale <strong>de</strong>s <strong>de</strong>ux ratios. De plus, ce test conjecture un taux <strong>de</strong> mutations<br />

synonymes stationnaire durant le processus d’évolution et l’absence <strong>de</strong> liaison génétique<br />

entre sites nucléotidiques (effet autostop négligé).<br />

10.3. Tests <strong>de</strong> polymorphisme visant à détecter un écart à la neutralité :<br />

tests <strong>de</strong> Tajima et Fu & Li<br />

Tajima (1989) propose un test statistique basé sur <strong>de</strong>s données <strong>de</strong> polymorphisme<br />

intra-population. La statistique D =<br />

("H^ #" S<br />

)<br />

1/ 2<br />

dépend <strong>de</strong> la différence entre <strong>de</strong>ux<br />

[ Var("H^ #" S<br />

)] estimateurs du thêta <strong>de</strong> Watterson (1975) ou diversité nucléotidique : "H^ le nombre moyen<br />

<strong>de</strong> différences entre les ! paires <strong>de</strong> séquences échantillonnées et θ S le nombre total <strong>de</strong> sites<br />

polymorphes observés ; pondérée par l’estimation <strong>de</strong> la variance <strong>de</strong> cette différence. Si le<br />

!<br />

polymorphisme génétique s’explique par la seule dérive (polymorphisme neutre), alors ces<br />

<strong>de</strong>ux paramètres sont égaux et le D <strong>de</strong> Tajima est nul. Au contraire, un allèle en fréquence<br />

rare entraînera une augmentation du nombre total <strong>de</strong> sites polymorphes mais aura un<br />

impact minime sur le nombre moyen <strong>de</strong> différences entre paires <strong>de</strong> séquences (effet<br />

sensible à la taille <strong>de</strong> l’échantillonnage). Un excès <strong>de</strong> variants rares engendrera ainsi un D<br />

<strong>de</strong> Tajima négatif alors qu’il sera positif <strong>pour</strong> un excès <strong>de</strong> variants intermédiaires. L’excès<br />

<strong>de</strong> variants rares indique généralement une sélection positive récente (un allèle avantageux<br />

a récemment envahi la population et <strong>de</strong>s variants neutres encore rares sont apparus par<br />

mutation), un effet autostop (mutations neutres liées à l’avantage ou <strong>de</strong> la sélection<br />

négative (<strong>de</strong>s allèles faiblement délétères ségrégent en faible fréquence) et correspond à un<br />

D négatif. A l’inverse, un excès <strong>de</strong> variants intermédiaires est généralement dûà <strong>de</strong> la<br />

sélection balancée (qui maintient les allèles à <strong>de</strong>s fréquences intermédiaires) en l’absence<br />

<strong>de</strong> recombinaison et se traduira par un D positif. D suit approximativement une loi <strong>de</strong><br />

distribution bêta. Cependant, plusieurs conditions doivent être vérifiées : (1) les séquences<br />

71


Chapitre 2<br />

d’ADN doivent être échantillonnées au hasard dans la population, (2) il n’y a pas <strong>de</strong><br />

phénomène d’autostop et surtout, (3) la population a atteint l’équilibre mutation / dérive<br />

<strong>de</strong>puis une longue pério<strong>de</strong> évolutive. En effet, ce test ne permet pas <strong>de</strong> faire la différence<br />

entre la sélection darwinienne et l’existence <strong>de</strong> changements démographiques passés.<br />

Ainsi, par exemple, un D <strong>de</strong> Tajima négatif peut être dû à la sélection mais également à un<br />

goulôt d’étranglement <strong>de</strong> la population suivi d’une phase d’expansion. De plus, ce test<br />

s’avère peu robuste <strong>pour</strong> un faible nombre <strong>de</strong> séquences échantillonnées. Par conséquent, il<br />

faut rester pru<strong>de</strong>nt lors <strong>de</strong> l’interprétation <strong>de</strong>s résultats. Fu et Li (1993) proposent une<br />

amélioration du test <strong>de</strong> Tajima. en estimant la diversité nucléotidique uniquement à partir<br />

<strong>de</strong>s mutations récemment apparues (singletons présents sur une seule séquence). Ce<br />

<strong>de</strong>rnier test apparaît donc beaucoup plus sensible aux mutations artéfactuelles. La<br />

comparaison <strong>de</strong> ces indices sur un gène <strong>pour</strong> plusieurs espèces et/ou plusieurs gènes d’une<br />

même espèce partageant la même zone géographique, telle que l’EPR, peut néanmoins<br />

favoriser l’hypothèse sélective ou démographique. En effet, un événement démographique<br />

provoqué par une modification <strong>de</strong> l’habitat (e.g. épanchement <strong>de</strong> lave détruisant la faune)<br />

doit théoriquement affecter l’ensemble <strong>de</strong>s gènes d’une espèce et la majorité <strong>de</strong>s espèces<br />

(Hudson et al 1987, Galtier et al 2000). Au contraire, la sélection n’affectera qu’un faible<br />

nombre <strong>de</strong> gènes et d’espèces. En conséquence, un D <strong>de</strong> Tajima négatif sur une même<br />

région géographique <strong>pour</strong> l’ensemble <strong>de</strong>s espèces et <strong>de</strong>s gènes posera l’hypothèse forte<br />

d’une expansion démographique (goulôt d’étranglement réduisant la diversité, suivi d’une<br />

accumulation <strong>de</strong> nouvelles mutations : excès <strong>de</strong> variants rares). Cette hypothèse sera<br />

d’autant plus forte que l’expansion supposée est simultanée <strong>pour</strong> l’ensemble <strong>de</strong> ces<br />

espèces.<br />

10.4. Dater une expansion démographique via le logiciel Fluctuate<br />

Le logiciel Fluctuate (Kuhner et al 1998) permet <strong>de</strong> dater le début <strong>de</strong> l’expansion<br />

démographique d’une population à partir <strong>de</strong> la structure généalogique d’un coalescent. En<br />

effet, en absence <strong>de</strong> sélection, une population en expansion doit présenter un coalescent en<br />

étoile, avec une séquence majoritaire ancestrale et <strong>de</strong> multiple séquences dérivées<br />

faiblement divergentes <strong>de</strong> cette <strong>de</strong>rnière. Le logiciel consiste à échantillonner un grand<br />

nombre <strong>de</strong> généalogies neutres dans l’univers <strong>de</strong>s possibles (par la métho<strong>de</strong> <strong>de</strong> Metropolis-<br />

Hasting utilisant <strong>de</strong>s chaînes <strong>de</strong> Markov sous procédure Monte Carlo, MCMC, Metropolis<br />

72


Chapitre 2<br />

et al 1953), à partir <strong>de</strong>s données <strong>de</strong> séquences fournies, une valeur initiale du paramètre<br />

populationnel θ (=2 Ne µ <strong>pour</strong> un gène mitochondrial et 4 Ne µ <strong>pour</strong> un gène nucléaire, où<br />

Ne correspond à la taille efficace <strong>de</strong> la population et µ au taux <strong>de</strong> mutations) et du<br />

paramètre <strong>de</strong> croissance exponentielle g. Une probabilité postérieure est ensuite attribuée à<br />

chacune <strong>de</strong> ces généalogies (Griffiths & Tavare 1994a, 1994b), la meilleure d’entre elles<br />

permettant <strong>de</strong> réévaluer conjointement les valeurs <strong>de</strong> θ et g.<br />

Cette approche suppose l’absence <strong>de</strong> sélection sur le locus considéré, l’absence <strong>de</strong><br />

recombinaison intra-locus, que la population soit fermée (pas <strong>de</strong> migration) et une<br />

expansion exponentielle <strong>de</strong> la population au cours du temps (t) à partir d’une population<br />

ancestrale <strong>de</strong> taille efficace Ne a (Kuhner et al 1998). Elle fait également appel au taux <strong>de</strong><br />

mutation du locus considéré en faisant l’hypothèse que ce taux <strong>de</strong> mutation est constant au<br />

cours du temps. La relation θ t = θ a e -gµt (où θ t et θ a correspon<strong>de</strong>nt à θ au temps t et à la<br />

population ancestrale, respectivement, Kuhner et al 1998) permet ainsi <strong>de</strong> corréler la taille<br />

<strong>de</strong> la population au temps et ainsi <strong>de</strong> déterminer le début <strong>de</strong> l’expansion démographique.<br />

Cette approche peut également être réalisée <strong>pour</strong> dater le début d’un déclin démographique<br />

(g est alors négatif).<br />

10.5. Proposition d’un nouveau test statistique visant à détecter <strong>de</strong>s locus<br />

sous sélection diversifiante entre régions géographiques<br />

Nous proposons un nouveau test multi-locus visant à détecter <strong>de</strong>s locus sous<br />

sélection diversifiante entre régions géographiques différentes. Ce test est basé sur la<br />

différence entre les thêtas <strong>de</strong> Watterson Δθ w calculés dans <strong>de</strong>s lignées présentes dans <strong>de</strong>s<br />

régions géographiques différentes (e.g. Nord et Sud). Les Δθ w (= θ w Nord - θ w Sud)<br />

observées sont ainsi calculées <strong>pour</strong> chaque gène et doivent être distribuées selon une loi<br />

normale sous l’hypothèse d’un modèle neutre d’isolement <strong>de</strong> ces régions géographiques.<br />

Les gènes présentant une valeur <strong>de</strong> Δθ w en marge <strong>de</strong> la distribution normale seraient ainsi<br />

<strong>de</strong> bons candidats <strong>pour</strong> détecter <strong>de</strong> la sélection diversifiante entre les régions<br />

géographiques considérées. Ce test peut être considéré indépendant <strong>de</strong>s taux <strong>de</strong> mutation<br />

<strong>de</strong> chaque locus car il est supposé que les gènes considérés présentent le même taux <strong>de</strong><br />

mutation dans ces <strong>de</strong>ux régions. Le logiciel MsBayes (Hickerson et al 2006b) peut être<br />

utilisé afin <strong>de</strong> simuler les valeurs <strong>de</strong> θ w dans chaque paire <strong>de</strong> lignées sous modèle neutre<br />

73


Chapitre 2<br />

(nécessité <strong>de</strong> préciser l’absence <strong>de</strong> goulôt d’étranglement dans MsBayes) et ainsi d’obtenir<br />

la distribution <strong>de</strong>s probabilités postérieures <strong>de</strong>s Δθ w utilisée <strong>pour</strong> détecter les Δθ w en marge<br />

<strong>de</strong> la distribution.<br />

Ce test a été utilisé <strong>pour</strong> les espèces A. pompejana et B. thermophilus mais n’a pas<br />

permis <strong>de</strong> détecter <strong>de</strong> locus « outlier », la totalité <strong>de</strong>s locus étudiés étant situés en milieu <strong>de</strong><br />

distribution. Il serait néanmoins intéressant <strong>de</strong> tester la puissance statistique <strong>de</strong> ce test en<br />

l’utilisant sur <strong>de</strong>s locus connus comme étant soumis à sélection diversifiante entre régions<br />

géographiques différentes.<br />

10.6. Discrimination entre balayage sélectif et goulôt d’étranglement via<br />

une approche <strong>de</strong> vraisemblance sur données multilocus : le logiciel<br />

Sweep_bott (Galtier et al 2000)<br />

Afin <strong>de</strong> discriminer l’effet d’une sélection positive (ou, par autostop moléculaire, <strong>de</strong><br />

balayage sélectif) d’un goulôt d’étranglement <strong>de</strong> la population, une métho<strong>de</strong> basée sur <strong>de</strong>s<br />

estimations <strong>de</strong> maximum <strong>de</strong> vraisemblance <strong>de</strong> coalescents à plusieurs locus sous<br />

l’hypothèse d’un modèle <strong>de</strong> réduction <strong>de</strong> la diversité génétique (logiciel Sweep_bott,<br />

Galtier et al 2000) a également été utilisée lors <strong>de</strong>s approches multigènes. L’idée est <strong>de</strong><br />

considérer qu’un goulôt d’étranglement se traduira par une réduction <strong>de</strong> polymorphisme à<br />

un temps T et avec une force S, ces paramètres étant i<strong>de</strong>ntiques entre tous les locus<br />

échantillonnés. À l’inverse, le temps et la force d’un balayage sélectif seront différents<br />

voire nuls selon le locus étudié. Cette métho<strong>de</strong> faisant l’hypothèse d’un modèle en nombre<br />

<strong>de</strong> sites infini (une mutation ne pouvant survenir qu’une seule fois sur un site donné), les<br />

sites s’écartant <strong>de</strong> ce modèle (i.e. site à plus <strong>de</strong> <strong>de</strong>ux états <strong>de</strong> caractère, les séquences<br />

présentant <strong>de</strong>s sites homoplasiques ayant déjà été éliminées <strong>pour</strong> obtenir un jeu <strong>de</strong> qualité<br />

4 à R m nul) sont supprimés <strong>pour</strong> les analyses utilisant sweep_bott. Nécessitant un codage<br />

binaire <strong>de</strong>s séquences avec 0 représentant l’état ancestral du site et 1 l’état dérivé, les états<br />

<strong>de</strong> caractère les plus fréquemment échantillonnés <strong>pour</strong> chaque site donné sont considérés<br />

ancestral (le second état <strong>de</strong> caractère échantillonné étant ainsi dérivé). Trois modèles sont<br />

définis:<br />

(1) Modèle M1 : modèle neutre (absence d’événement démographique et sélectif),<br />

comptant p paramètres correspondant aux valeurs <strong>de</strong> θ <strong>pour</strong> chacun <strong>de</strong>s p locus<br />

74


Chapitre 2<br />

(2) Modèle M2 : modèle « goulôt d’étranglement » (bottleneck). Il compte p+2<br />

paramètres : p valeurs <strong>de</strong> θ, S et T. Pour un événement démographique, S et T<br />

sont les mêmes <strong>pour</strong> l’ensemble <strong>de</strong>s locus.<br />

(3) Modèle M3 : modèle <strong>de</strong> « balayage sélectif » (selective sweep). Il compte 3p<br />

paramètres. En effet, chacun <strong>de</strong>s p locus compte une valeur <strong>de</strong> θ, un S et un T<br />

différents.<br />

Pour chaque modèle, une valeur <strong>de</strong> maximum <strong>de</strong> vraisemblance est calculée (selon<br />

Griffiths & Tavare 1994a, 1994b). Un test <strong>de</strong> différences relatives <strong>de</strong> vraisemblance<br />

(likelihood ratio test, LRT = 2(L 2 -L 1 ) distribué selon une loi du χ 2 , où L 1 et L 2<br />

correspon<strong>de</strong>nt aux maximum <strong>de</strong> vraisemblance <strong>de</strong>s modèles considérés) effectué entre<br />

chaque paire <strong>de</strong> modèles permet ainsi <strong>de</strong> définir le modèle le plus probable. Si le modèle<br />

M2 est favorisé (significativement différent <strong>de</strong> M1), alors l’ensemble <strong>de</strong>s locus doit<br />

présenter un écart à la neutralité (LRT par locus significatif). Au contraire, si le modèle M3<br />

est favorisé (significativement différent <strong>de</strong> M1 et M2), les locus sous balayage sélectif<br />

doivent montrer un écart à la neutralité, séparément. Dans le cas où, un locus sous<br />

balayage sélectif s’additionne à un goulôt d’étranglement, on s’attend à ce que : (1) les<br />

modèles M2 et M3 soient tous les <strong>de</strong>ux significativement différents du modèle M1, (2) le<br />

modèle M3 soit favorisé vis-à-vis du modèle M2 (au moins un locus présentant un T et un<br />

S différents <strong>de</strong>s autres locus), (3) tous les locus soient significativement différents du<br />

modèle M1 mais, que le locus (ou les locus) sous balayage sélectif présentent un LRT plus<br />

fort que le LRT <strong>de</strong>s locus non soumis à balayage.<br />

75


Chapitre 2<br />

76


Chapitre 3<br />

Chapitre 3 : Phylogéographie comparée <strong>de</strong>s<br />

espèces hydrothermales <strong>de</strong> la dorsale du<br />

Pacifique oriental à l’ai<strong>de</strong> du gène mitochondrial<br />

<strong>de</strong> la Cytochrome Oxydase I<br />

1. Introduction à l’approche <strong>de</strong> phylogéographie comparée <strong>de</strong>s espèces<br />

Jusqu’à aujourd’hui, la plupart <strong>de</strong>s étu<strong>de</strong>s phylogéographique ont été menées sur<br />

une espèce à la fois et visaient à mettre en évi<strong>de</strong>nce les processus ayant conduit à une<br />

différenciation génétique avec monophylie réciproque (divergence fixée entre lignées<br />

soeurs). Néanmoins, si la monophylie réciproque (Fig. 3.1) entre lignées constitue un cas<br />

extrême <strong>de</strong> différenciation génétique, pouvant mener à la spéciation, le processus<br />

d’isolement génétique est progressif. Les généalogies <strong>de</strong> gènes vont ainsi passer par une<br />

phase <strong>de</strong> tri <strong>de</strong>s lignées allant <strong>de</strong> la polyphylie (absence d’isolement génétique) à <strong>de</strong> la<br />

paraphylie (isolement génétique incomplet) avant <strong>de</strong> présenter, potentiellement, une<br />

monophylie réciproque (Nosil et al 2009, Fig. 3.1).<br />

Pour qu’il y ait monophylie réciproque, le polymorphisme ancestral doit être éliminé<br />

par dérive génétique et/ou sélection naturelle. Le temps nécessaire <strong>pour</strong> atteindre ce sta<strong>de</strong><br />

<strong>de</strong> différenciation dépendra donc <strong>de</strong> la taille efficace <strong>de</strong>s populations (Avise 2000). Ainsi,<br />

la probabilité d’observer une monophylie réciproque lors d’un isolement <strong>de</strong> <strong>de</strong>ux<br />

populations par une barrière à la dispersion, augmente au cours du temps et s’avère<br />

d’autant plus forte que la taille efficace <strong>de</strong> la population est petite (Fig. 3.2).<br />

77


Chapitre 3<br />

Gradient d’isolement<br />

génétique<br />

Initiation<br />

Isolement total<br />

Isolement<br />

génétique <strong>de</strong>s<br />

individus<br />

Coalescent<br />

Fig. 3.1 Etapes <strong>de</strong> mise en place d’un isolement génétique complet et <strong>de</strong>s coalescents théoriques<br />

associés à ces étapes. Adaptée <strong>de</strong> Nosil et al (2009).<br />

Les individus sont représentés par <strong>de</strong>s cercles, la couleur indique le génotype. Les individus bicolores<br />

correspon<strong>de</strong>nt à <strong>de</strong>s individus hybri<strong>de</strong>s.<br />

0 Ne 2Ne 3Ne 4Ne<br />

Nombre <strong>de</strong> générations <strong>de</strong>puis la divergence<br />

Fig. 3.2 Probabilité d’obtenir un coalescent polyphylétique, paraphylétique ou en monophylie<br />

réciproque en fonction du temps écoulé <strong>de</strong>puis le début <strong>de</strong> la divergence en allopatrie <strong>de</strong> <strong>de</strong>ux<br />

populations. Adaptée <strong>de</strong> Neigel & Avise (1986)<br />

La taille efficace Ne est supposée i<strong>de</strong>ntique <strong>pour</strong> les <strong>de</strong>ux populations et constante au cours du temps<br />

78


Chapitre 3<br />

De plus, la monophylie réciproque n’est pas nécessairement immuable. En effet, si<br />

l’isolement reproducteur n’est pas complet (spéciation non atteinte), une remise en contact<br />

secondaire entre lignées géographiquement isolées, par exemple, peut engendrer <strong>de</strong><br />

l’hybridation entre ces lignées, donnant ainsi lieu à un coalescent para- voire<br />

polyphylétique. Ainsi, un coalescent para/poly-phylétique peut être dû soit à une rétention<br />

<strong>de</strong> polymorphisme ancestrale (Fig. 3.3.A), soit à une remise en contact secondaire après un<br />

relâchement <strong>de</strong> la barrière aux flux <strong>de</strong> gène (Fig. 3.3.B). Dans les 2 cas, cela se traduit par<br />

un mélange d’allèles au niveau <strong>de</strong>s 2 lignées dans les populations échantillonnées mais<br />

dans lequel la coalescence <strong>de</strong> l’allèle « introduit » est généralement profon<strong>de</strong> s’il est<br />

associé à un polymorphisme ancestral non trié, ou terminale si le contact secondaire est<br />

récent.<br />

A Site 1 Site 2 B Site 1 Site 2<br />

Temps en générations<br />

1<br />

1<br />

10 1<br />

10 1<br />

10 4<br />

10 4<br />

10 7 10 7<br />

Barrière aux flux <strong>de</strong> gène<br />

Barrière aux flux <strong>de</strong> gène<br />

Fig. 3.3 Deux scénarii alternatifs <strong>pour</strong> expliquer un coalescent para/poly-phylétique entre<br />

<strong>de</strong>ux sites géographiquement distincts. Adaptée <strong>de</strong> Avise (2000)<br />

A, rétention <strong>de</strong> polymorphisme ancestral ; B, remise en contact secondaire.<br />

Avise (1987) suppose que les capacités <strong>de</strong> dispersion jouent un rôle majeur dans la<br />

structure géographique <strong>de</strong>s lignées. Chez les invertébrés marins, la présence d’une phase<br />

larvaire planctonique, sa durée <strong>de</strong> vie et la possibilité <strong>de</strong> migration <strong>de</strong>s individus adultes sont<br />

autant <strong>de</strong> traits d’histoire <strong>de</strong> vie qui favorisent la dispersion. Ils interagissent avec les<br />

79


Chapitre 3<br />

contraintes environnementales et peuvent ainsi permettre ou non <strong>de</strong> franchir <strong>de</strong>s barrières à la<br />

dispersion. Généralement, plus les traits d’histoire <strong>de</strong> vie favorisent la dispersion et plus la<br />

structure génétique observée est faible (Palumbi 1994 ; Féral 2002 ; Kinlan & Gaines 2003).<br />

L’analyse <strong>de</strong> séquences du gène mitochondrial Cytochrome Oxydase I sur plusieurs espèces<br />

<strong>de</strong> crépidules (gastéropo<strong>de</strong>s invasifs) le long <strong>de</strong> la côte est <strong>de</strong> l’Amérique du Nord, a montré<br />

que les espèces à développement direct présentaient une structure génétique plus marquée que<br />

celles à développement larvaire planctotrophe (Collin 2001). Wilke & Davis (2000)<br />

considèrent que la différenciation génétique observée chez le gastéropo<strong>de</strong> Hydrobia ventrosa<br />

(développement direct) est plus élevée que chez H. ulvae (1-3 jours <strong>de</strong> sta<strong>de</strong> larvaire<br />

planctonique), et <strong>pour</strong>rait être liée à <strong>de</strong>s différences d’habitat entre les populations d’adultes<br />

plutôt qu’au mo<strong>de</strong> <strong>de</strong> développement lui-même. Rolán-Alvarez et al (1995) suggèrent, quant à<br />

eux, que les différences <strong>de</strong> structure génétique entre <strong>de</strong>ux espèces <strong>de</strong> littorines (Littorina<br />

mariae et L. obtusata) à développement direct seraient dues à un temps <strong>de</strong> génération<br />

différent entre ces espèces. De nombreux critères (cycle <strong>de</strong> vie, morphologie, physiologie,<br />

comportement, etc) influençant les capacités <strong>de</strong> dispersion <strong>de</strong>s espèces, mais également leur<br />

potentiel d’adaptation à un habitat donné, entrent ainsi en jeu et affectent la topologie <strong>de</strong>s<br />

coalescents. Lors d’une étu<strong>de</strong> monospécifique, l’absence <strong>de</strong> monophylie réciproque entre<br />

populations issues <strong>de</strong> sites géographiquement distincts <strong>pour</strong>rait alors signifier :<br />

(1) l’absence <strong>de</strong> barrière à la dispersion entre ces sites<br />

(2) l’existence d’une barrière à la dispersion mais qui ne constituerait pas un<br />

obstacle <strong>pour</strong>, celle-ci présentant potentiellement <strong>de</strong> meilleures capacités <strong>de</strong> dispersion que<br />

les autres espèces plus structurées<br />

(3) l’existence d’une barrière ayant anciennement impacté cette espèce mais plus<br />

récemment perméable aux flux <strong>de</strong> gènes après la remise en contact secondaire <strong>de</strong>s lignées<br />

Afin <strong>de</strong> mieux connaître les causes <strong>de</strong> la distribution géographique <strong>de</strong>s lignées<br />

intraspécifiques d’une espèce, il est intéressant <strong>de</strong> comparer les patrons<br />

phylogéographiques <strong>de</strong> cette espèce à ceux d’autres espèces présentant une distribution<br />

géographique similaire. Ainsi, les concordances/discordances phylogéographiques entre<br />

espèces informeront sur l’impact d’une barrière aux flux <strong>de</strong> gènes et donc sur l’intensité ou<br />

sur l’âge <strong>de</strong> la barrière responsable <strong>de</strong> la structure géographique observée (Zink 2002).<br />

Cette approche comparative apporte ainsi <strong>de</strong>s informations précieuses sur les facteurs<br />

ayant façonné la distribution géographique <strong>de</strong>s lignées, et peut permettre <strong>de</strong> déduire la<br />

nature <strong>de</strong> cette barrière (e.g. géologique, courantologie, distance à parcourir…). De plus, la<br />

phylogéographie comparée entre espèces peut également, dans certains cas, permettre <strong>de</strong><br />

déterminer les traits d’histoire <strong>de</strong> vie spécifiquement responsable <strong>de</strong>s dissemblances <strong>de</strong><br />

80


Chapitre 3<br />

comportement <strong>de</strong>s espèces face à cette barrière (Bermingham & Avise 1986, Bermingham<br />

1998). De plus, l’analyse conjointe <strong>de</strong> plusieurs taxons issus <strong>de</strong> régions ayant <strong>de</strong>s histoires<br />

communes telles que celles associées à la fermeture <strong>de</strong> l’Isthme <strong>de</strong> Panama ou <strong>de</strong><br />

l’ouverture du détroit <strong>de</strong> Béring peut fournir <strong>de</strong>s informations précieuses sur les logiques<br />

d’extinction totale ou partielle <strong>de</strong> certaines lignées voire d’inférer <strong>de</strong>s scénarios sur la<br />

recolonisation d’une région/environnement par la lignée ayant survécu dans l’autre<br />

région/environnement (Cunningham & Collins 1998).<br />

Dans ce chapitre, nous avons effectué une approche <strong>de</strong> phylogéographie comparée<br />

<strong>de</strong> sept espèces hydrothermales profon<strong>de</strong>s <strong>de</strong> la dorsale du Pacifique oriental (EPR) en<br />

utilisant le gène mitochondrial <strong>de</strong> la Cytochrome Oxydase I (mtCOI). Ce chapitre<br />

correspond à un article publié dans la revue Molecular Ecology, présenté après un résumé<br />

en français. Il a <strong>pour</strong> objectif <strong>de</strong> tester l’hypothèse d’une barrière aux flux <strong>de</strong> gènes au<br />

niveau <strong>de</strong> la zone équatoriale <strong>de</strong> l’EPR pouvant avoir donné lieu à un événement vicariant<br />

majeur ayant séparé les faunes hydrothermales du Pacifique est. Cette étu<strong>de</strong> <strong>de</strong> ce fait<br />

<strong>pour</strong>rait permettre <strong>de</strong> vali<strong>de</strong>r l’hypothèse <strong>de</strong> 2 sous-unités biogéographiques proposées par<br />

Bachraty et al (2009) et éventuellement <strong>de</strong> dater cette séparation. Le choix <strong>de</strong> cette zone a<br />

été guidée par les travaux d’Hurtado et al (2004) qui soulignaient la possibilité d’une<br />

barrière au niveau du point triple <strong>de</strong>s Galapagos chez les polychètes Alvinella pompejana<br />

et Tevnia jerichonana sur la base <strong>de</strong> monophilies réciproques <strong>de</strong> part et d’autre <strong>de</strong> cette<br />

barrière.<br />

2. Résumé <strong>de</strong> l’article <strong>de</strong> phylogéographie comparée entre espèces<br />

Pour effectuer notre approche <strong>de</strong> phylogéographie comparée entre espèces, nous<br />

avons choisi d’utiliser <strong>de</strong>s espèces appartenant à <strong>de</strong>s familles différentes (gastéropo<strong>de</strong>s :<br />

Lepetodrilus elevatus, L. ovalis, Eulepetopsis vitrea, polychètes : Alvinella pompejana,<br />

Hesiolyra bergi, Branchipolynoe symmytilida, et un bivalve : Bathymodiolus thermophilus,<br />

Fig. 3.4) afin d’obtenir une vue d’ensemble <strong>de</strong>s distributions phylogéographiques<br />

d’espèces présentant <strong>de</strong>s traits d’histoire <strong>de</strong> vie contrastés. Ces espèces sont relativement<br />

abondantes sur l’EPR, permettant l’échantillonnage nécessaire à une étu<strong>de</strong> <strong>de</strong> populations,<br />

et endémiques <strong>de</strong> l’EPR. Ces espèces ont été échantillonnées sur neuf champs<br />

81


Chapitre 3<br />

hydrothermaux répartis <strong>de</strong> 21°N à 21°S. Deux à trois cents individus par espèce ont été<br />

séquencés sur un fragment du gène mtCOI.<br />

Malgré les disparités biologiques entre ces espèces, les réseaux d’haplotypes du gène<br />

mtCOI (Fig. 3.5) révèlent <strong>de</strong>s topologies similaires avec <strong>de</strong>ux cla<strong>de</strong>s divergents, à<br />

l’exception du polychète commensal <strong>de</strong>s moules hydrothermales, B. symmytilida. Cette<br />

<strong>de</strong>rnière espèce montre, en effet, un réseau plus complexe suggérant une diversification<br />

plus ancienne et l’absence <strong>de</strong> co-évolution entre ce polychète et son hôte.<br />

Le polychaete A. pompejana et le gastéropo<strong>de</strong> E. vitrea présentent exactement la<br />

même architecture <strong>de</strong> réseau d’haplotypes (Fig. 3.5), avec <strong>de</strong>ux cla<strong>de</strong>s distints (divergence<br />

<strong>de</strong> 1% <strong>pour</strong> A. pompejana et 0,9% <strong>pour</strong> E. vitrea) séparés <strong>de</strong> part et d’autre <strong>de</strong> l’Equateur.<br />

L. elevatus compte <strong>de</strong>ux cla<strong>de</strong>s A et B divergents <strong>de</strong> 6,5% correspondant aux <strong>de</strong>ux<br />

espèces cryptiques définies par Matabos et al (2008b) distribués respectivement <strong>de</strong> 21°N à<br />

9°50’N (cla<strong>de</strong> A) et <strong>de</strong> 9°50’N à 21°S (cla<strong>de</strong> B). Quelques individus du cla<strong>de</strong> B ayant été<br />

également échantillonnés à 13°N. La population <strong>de</strong> 21°N est composée uniquement <strong>de</strong><br />

l’haplotype 1. La population <strong>de</strong> 9°50’N diffèrent légèrement <strong>de</strong>s autres populations du Sud<br />

par la présence <strong>de</strong> l’haplotype 4 trouvé en plus forte fréquence sur ce site, expliquant la<br />

significativitédu φ st calculé au sein du cla<strong>de</strong> B entre cette population et celles situées plus<br />

au Sud.<br />

Les <strong>de</strong>ux cla<strong>de</strong>s détectés chez L. ovalis (2% <strong>de</strong> divergence) forment un cline <strong>de</strong> 14°S<br />

à 21°S, alors que les champs hydrothermaux 9°50’N et 7°25’S ne sont composés que<br />

d’haplotypes du cla<strong>de</strong> A. Le bivalve B. thermophilus montre, à l’inverse, un cline <strong>de</strong> ces<br />

cla<strong>de</strong>s divergents au Nord (13°N à 7°25’S) et <strong>de</strong>s séquences du cla<strong>de</strong> B uniquement au Sud<br />

(14°S à 21°S).<br />

H. bergi montre un réseau d’haplotype en étoile avec un haplotype central distribué<br />

sur l’ensemble <strong>de</strong> l’EPR et quelques individus du Sud présentant une divergence <strong>de</strong> 1%.<br />

La distribution géographique <strong>de</strong>s haplotypes met ainsi en exergue la présence d’une zone<br />

<strong>de</strong> transition entre le nord et le sud <strong>de</strong> la dorsale. Cette zone <strong>de</strong> transition est confirmée par<br />

l’augmentation <strong>de</strong>s valeurs <strong>de</strong> différenciation génétique (φ st ) entre 17°S et l’équateur.<br />

82


Chapitre 3<br />

Bathymodiolus thermophilus<br />

(bivalve formant <strong>de</strong> véritables moulières)<br />

Alvinella pompejana<br />

(polychète tubicole)<br />

Hesiolyra bergi<br />

(polychète libre, proche <strong>de</strong>s<br />

tubes d’A. pompejana)<br />

Branchipolynoe symmytilida<br />

(polychète commensal <strong>de</strong>s<br />

Bathymodiolus)<br />

Eulepetopsis vitrea<br />

(gastéropo<strong>de</strong> associé au<br />

pôle froid <strong>de</strong>s sources<br />

hydrothermales)<br />

Lepetodrilus elevatus (à gauche) et L. ovalis (à droite), gastéropo<strong>de</strong>s<br />

abondants sur <strong>de</strong>s roches, Riftia pachyptila (au centre) et Bathymodiolus<br />

Fig. 3.4 Photographies <strong>de</strong>s espèces étudiées par l’approche <strong>de</strong> phylogéographie comparée<br />

sur le mtCOI. Tirées <strong>de</strong> Desbruyères et al (2006b)<br />

83


Chapitre 3<br />

Fig. 3.5 Réseaux d’haplotypes en Median Joining et distributions <strong>de</strong>s fréquences haplotypiques du<br />

gène mtCOI <strong>pour</strong> les sept espèces étudiées.<br />

Sur les réseaux, la taille <strong>de</strong>s cercles d’haplotypes et les connexions entre haplotypes sont proportionnelles<br />

au nombre <strong>de</strong> séquences et <strong>de</strong> pas mutationnels, respectivement. A et B représentent les <strong>de</strong>ux cla<strong>de</strong>s<br />

divergents <strong>pour</strong> chaque espèce.<br />

Sur les distributions d’haplotypes, les haplotypes partagés ayant une fréquence > 2% au sein du cla<strong>de</strong><br />

considéré sont colorés. Les haplotypes privés et haplotypes partagés en fréquence < 2% au sein du cla<strong>de</strong><br />

considéré sont représentés en blanc. Bt, Bathymodiolus thermophilus ; Ap, Alvinella pompejana ; Hb,<br />

Hesiolyra bergi ; Bs, Branchipolynoe symmytilida ; Ev, Eulepetopsis vitrea ; Lo, Lepetodrilus ovalis ; Le,<br />

Lepetodrilus elevatus.<br />

84


Chapitre 3<br />

L’hypothèse d’horloge moléculaire ayant été testée et validée sur les jeux <strong>de</strong><br />

données, nous avons ensuite cherché à déterminer si les divergences observées entre les<br />

<strong>de</strong>ux cla<strong>de</strong>s <strong>de</strong> chaque espèce (pouvant correspondre aux lignées ayant divergé <strong>de</strong> part et<br />

d’autre d’une même barrière entre nord et sud EPR) étaient issues d’un même événement<br />

<strong>de</strong> vicariance. Pour cela, nous avons utilisé une nouvelle approche basée sur une métho<strong>de</strong><br />

d’ « Approximate Bayesian Computation » (métho<strong>de</strong> dite ABC) à l’ai<strong>de</strong> du logiciel<br />

MsBayes (Hickerson et al 2007, voir le chapitre 2 <strong>pour</strong> plus d’explication). Cette analyse a<br />

permis <strong>de</strong> mettre en évi<strong>de</strong>nce la présence <strong>de</strong> <strong>de</strong>ux évènements <strong>de</strong> divergence datés<br />

respectivement à 11,6 Ma <strong>pour</strong> L. elevatus uniquement et, correspondant aux <strong>de</strong>ux espèces<br />

cryptiques découvertes par Matabos et al (2008b), et à 1,3 Ma <strong>pour</strong> les six autres espèces.<br />

L’observation <strong>de</strong> la divergence entre les haplotypes 3 (sud EPR) et 4 (principalement à<br />

9°50’N) chez l’espèce cryptique L. elevatus majoritairement trouvée <strong>de</strong> 9°50’N à 21°33’S,<br />

<strong>pour</strong>rait correspondre également à l’événement <strong>de</strong> divergence daté à 1,3 Ma. Ainsi, L.<br />

elevatus aurait à la fois subi un isolement allopatrique ayant donné lieu aux <strong>de</strong>ux espèces<br />

cryptiques et à la zone <strong>de</strong> contact à 9°50’N, mais également un isolement concomitant à<br />

celui <strong>de</strong>s autres espèces à 1,3 Ma entre 9°50’N et le sud <strong>de</strong> l’EPR (haplotype 4 plus<br />

fréquent à 9°50’N).<br />

Il y a 12,5-11 Ma, la Mathematician Ridge, située au Nord, a subi d’importantes<br />

réorientations magnétiques et bathymétriques, provoquant la formation d’un système <strong>de</strong><br />

dorsales parallèles (microplate Mathematician) avec la Mathematician Ridge à l’ouest et la<br />

Moctezuma Trough à l’est (Mammerickx & Kligtgord 1982, voir chapitre 1). Ce n’est<br />

qu’environ 4-5 Ma plus tard (7-5 Ma) que la Moctezuma Trough (survivante <strong>de</strong>s 2 dorsales<br />

après fossilisation <strong>de</strong> la Mathematician Ridge) et l’oEPR aurait fusionné <strong>pour</strong> former<br />

l’EPR actuelle. Cette réorganisation <strong>de</strong> la dorsale au Nord avec la formation temporaire <strong>de</strong><br />

la microplaque Mathematician <strong>pour</strong>rait avoir été à l’origine <strong>de</strong> la spéciation chez L.<br />

elevatus, il y a 11,6 Ma. Concernant le <strong>de</strong>uxième événement <strong>de</strong> vicariance ayant affecté<br />

l’ensemble <strong>de</strong>s espèces aux alentours <strong>de</strong> l’équateur, il <strong>pour</strong>rait être lié à la formation <strong>de</strong><br />

plusieurs failles transformantes entre 9°50’N et 17°25’S (i.e. système <strong>de</strong> fracture<br />

Quebrada/Discovery/Gofar, ou failles transformantes Wilkes et Garrett) qui ont fracturé<br />

l’axe <strong>de</strong> la dorsale en multiples fragments il y a 1-2 Ma (Kureth & Rea 1981, Naar & Hey<br />

1989, Francheteau et al 1990). Cette barrière aurait pu être renforcée par une discontinuité<br />

<strong>de</strong>s courants profonds dans cette zone (cf Reid 1997, Hurtado et al 2004).<br />

85


Chapitre 3<br />

La formation lente et progressive d’une barrière tectonique associée à <strong>de</strong>s<br />

caractéristiques <strong>de</strong> dispersion et/ou d’adaptation contrastés chez les espèces étudiées,<br />

<strong>pour</strong>raient expliquer que la barrière aux flux <strong>de</strong> gènes se situe à <strong>de</strong>s positions<br />

géographiques légèrement différentes entre espèces (Knowlton & Weigt 1998). Il est<br />

extrêmement difficile <strong>de</strong> détecter un critère biologique et/ou écologique unique pouvant<br />

expliquer ces différences. A. pompejana et H. bergi partagent le même habitat (proche du<br />

flui<strong>de</strong>, H. bergi se déplaçant autour voire dans les tubes d’A. pompejana) et, présentent<br />

néanmoins <strong>de</strong>s distributions d’haplotypes très contrastées. Ces distributions sont marquées<br />

par une barrière hermétique à l’équateur <strong>pour</strong> A. pompejana, et par la survivance d’une<br />

lignée sud divergente d’H. bergi très rarement échantillonnée au-<strong>de</strong>là <strong>de</strong> 17°25S. Le<br />

microenvironnement seul ne semble donc pas jouer <strong>de</strong> rôle dans la différence <strong>de</strong><br />

distribution <strong>de</strong> ces espèces. Les caractères dispersifs <strong>de</strong>s espèces hydrothermales restent<br />

relativement peu connus mais ne semble pas expliquer non plus ces différences. La taille<br />

<strong>de</strong>s œufs, leur contenu en vitellus ou le comportement <strong>de</strong>s larves peuvent influencer la<br />

migration verticale <strong>de</strong>s propagules (Mullineaux et al 2005, Arellano & Young 2009). La<br />

force et le sens <strong>de</strong>s courants océaniques peuvent varier en fonction <strong>de</strong> la profon<strong>de</strong>ur. Par<br />

conséquent, la migration verticale <strong>de</strong>s propagules entre espèces <strong>pour</strong>rait jouer un rôle dans<br />

les dissemblances observées dans le franchissement <strong>de</strong> la barrière. Cela <strong>pour</strong>rait également<br />

expliquer les résultats obtenus à l’ai<strong>de</strong> du logiciel Migrate 3.0.3 (Beerli & Felsenstein<br />

1999, 2001), avec H. bergi présentant une migration orientée <strong>de</strong> la lignée mitochondriale la<br />

plus fréquente du Nord vers le Sud alors que celles <strong>de</strong> L. elevatus, L. ovalis, B.<br />

thermophilus et B. symmytilida auraient plutôt migré du Sud vers le Nord. Des étu<strong>de</strong>s<br />

complémentaires sur les capacités <strong>de</strong> dispersion <strong>de</strong>s espèces <strong>pour</strong>raient permettre <strong>de</strong> mieux<br />

comprendre ces dissemblances.<br />

La forme en étoile <strong>de</strong>s réseaux d’haplotypes ainsi que les D <strong>de</strong> Tajima<br />

significativement négatifs au Sud <strong>de</strong> l’EPR <strong>pour</strong> l’ensemble <strong>de</strong>s espèces étudiées,<br />

suggèrent une expansion démographique au Sud <strong>de</strong> la dorsale. Cette expansion<br />

démographique serait concomitante entre toutes les espèces (hormis B. symmytilida) et<br />

datée <strong>de</strong> moins <strong>de</strong> 500 000 ans. Cette expansion démographique <strong>pour</strong>rait être due (1) à une<br />

colonisation récente <strong>de</strong> l’EPR Sud par <strong>de</strong>s effets fondateurs, en provenance, par exemple,<br />

<strong>de</strong> sites situés au Sud <strong>de</strong> la microplaque <strong>de</strong> Pâques (23°S), ou (2) à la fréquence plus élevée<br />

d’éruptions massives avec ensevelissement <strong>de</strong> la faune au Sud <strong>de</strong> l’EPR (fréquentes dans<br />

cette région, e.g. Haymon et al 1993, Cowen et al 2007). Ces <strong>de</strong>rnières auraient alors<br />

détruit massivement la faune et engendré <strong>de</strong>s goulôts d’étranglement plus aigus.<br />

86


Chapitre 3<br />

3. Article : phylogéographie comparée entre espèces sur le gène mtCOI<br />

87


Chapitre 3<br />

88


Chapitre 3<br />

89


Chapitre 3<br />

90


Chapitre 3<br />

91


Chapitre 3<br />

92


Chapitre 3<br />

93


Chapitre 3<br />

94


Chapitre 3<br />

95


Chapitre 3<br />

96


Chapitre 3<br />

97


Chapitre 3<br />

98


Chapitre 3<br />

99


Chapitre 3<br />

100


Chapitre 3<br />

101


Chapitre 3<br />

4. Intérêts et limites <strong>de</strong> l’approche <strong>de</strong> phylogéographie comparée<br />

Les approches <strong>de</strong> phylogéographie monospécifique permettent <strong>de</strong> révéler <strong>de</strong>s<br />

barrières aux flux <strong>de</strong> gènes ayant impacté l’espèce étudiée et <strong>de</strong> poser <strong>de</strong>s hypothèses<br />

quant à l’histoire évolutive <strong>de</strong> cette espèce. Les apports <strong>de</strong> la phylogéographie comparée<br />

ajoute une dimension supplémentaire en détectant <strong>de</strong>s changements environnementaux<br />

majeurs (cataclysmes tels qu’une éruption volcanique majeure, changements climatiques,<br />

fermeture ou ouverture d’un passage entre 2 régions) ayant impactés l’ensemble <strong>de</strong> la<br />

faune et la flore associée, tant du point <strong>de</strong> vue <strong>de</strong> la vicariance que <strong>de</strong> l’histoire<br />

démographique.<br />

4.1. Apports <strong>de</strong> la phylogéographie comparée entre espèces dans l’étu<strong>de</strong><br />

<strong>de</strong>s processus <strong>de</strong> vicariance<br />

L’estimation du temps <strong>de</strong> divergence T entre <strong>de</strong>ux lignées répond à une équation, en<br />

apparence, simple : T = D/2r où D correspond à la divergence génétique observée entre ces<br />

lignées et r, au taux <strong>de</strong> substitution du gène considéré. Il semble alors aisé, <strong>pour</strong> une espèce<br />

donnée, d’estimer la date d’isolement <strong>de</strong> <strong>de</strong>ux lignées monophylétiques et d’en rechercher<br />

la cause dans l’histoire <strong>de</strong> l’espèce (e.g. causes environnementales associées à la mise en<br />

place <strong>de</strong> la barrière). Néanmoins, cette corrélation entre divergence génétique et mise en<br />

place d’une barrière géographique pose en réalité <strong>de</strong> nombreux problèmes (e.g. Rambaut &<br />

Bromham 1998, Welch et al 2005).<br />

Tout d’abord, la corrélation temps <strong>de</strong> divergence/divergence génétique nécessite<br />

l’accès au taux <strong>de</strong> substitution du gène considéré. L’estimation d’un taux <strong>de</strong> substitution<br />

s’effectue généralement en comparant la divergence entre un ADN ancien issu <strong>de</strong> fossiles<br />

datés et un ADN actuel d’un même taxon (Shapiro et al 2004 : Rise and fall of the<br />

Beringian steppe bison, Science) ou par le biais <strong>de</strong> la divergence génétique entre lignées<br />

sœurs situées <strong>de</strong> part et d’autre d’une barrière géographique bien datée géologiquement<br />

(Knowlton & Weigt 1998, Chevaldonné et al 2002). Il peut être également obtenu <strong>de</strong> façon<br />

expérimental en suivant l’accumulation <strong>de</strong>s nouvelles mutations entre les lignées<br />

généalogiques d’une espèce donnée sur plusieurs générations (Holwell et al 2003).<br />

Néanmoins, cette technique est sujette à caution par le fait que les mutations délétères sont<br />

102


Chapitre 3<br />

purgées plus rapi<strong>de</strong>ment dans la divergence que dans le polymorphisme, donnant lieu à <strong>de</strong>s<br />

taux <strong>de</strong> mutation nettement plus élevée lorsqu’estimée par cette approche (Emerson 2007).<br />

Une fois estimé par l’intermédiaire d’une datation géologique précise, ce taux <strong>de</strong><br />

substitution <strong>pour</strong>rait alors être utilisé <strong>pour</strong> estimer un temps <strong>de</strong> divergence entre lignées<br />

<strong>pour</strong> une autre zone géographique. Néanmoins, le taux <strong>de</strong> substitution est souvent<br />

variable, non seulement selon l’espèce étudiée, mais aussi au cours du temps <strong>de</strong> manière<br />

indépendante dans chaque lignée évolutive (<strong>de</strong>ux lignées actuelles pouvant avoir évolué<br />

selon <strong>de</strong>s taux <strong>de</strong> substitution différents à partir d’une même lignée mère : e.g. Ho 2009).<br />

Afin <strong>de</strong> corréler la divergence génétique observée au temps <strong>de</strong> divergence, il est donc<br />

nécessaire <strong>de</strong> vérifier l’hypothèse d’horloge moléculaire (Zuckerkandl & Pauling 1965)<br />

qui suppose un taux <strong>de</strong> substitution constant au cours du temps et entre lignées évolutives.<br />

Un écart à la neutralité du gène considéré (balayage sélectif) peut également<br />

engendrer une mauvaise estimation du temps <strong>de</strong> divergence. En effet, la réduction du<br />

polymorphisme liée à <strong>de</strong> la sélection purifiante sur le gène considéré, engendrera ainsi une<br />

sous-estimation <strong>de</strong> la divergence entre cla<strong>de</strong>s par rapport à un coalescent neutre (voir<br />

l’introduction générale partie 3.2). La divergence estimée au gène sous sélection <strong>pour</strong>rait<br />

alors ne pas correspondre à l’événement ayant donné lieu à l’isolement <strong>de</strong>s espèces ou <strong>de</strong>s<br />

lignées intra-spécifiques.<br />

Sous les hypothèses <strong>de</strong> l’horloge moléculaire et <strong>de</strong> l’accumulation neutre <strong>de</strong>s<br />

mutations, la divergence génétique entre cla<strong>de</strong>s peut alors permettre <strong>de</strong> dater l’événement<br />

d’isolement <strong>de</strong> ces cla<strong>de</strong>s. Si ces cla<strong>de</strong>s sont sub-divisés géographiquement, alors cette<br />

datation peut être reliée à l’existence d’une potentielle barrière géographique aux flux <strong>de</strong><br />

gènes. Une approche phylogéographique espèce par espèce permet, en comparant les<br />

différentes dates obtenues, d’appréhen<strong>de</strong>r si ces espèces présentent <strong>de</strong>s dates <strong>de</strong> divergence<br />

proches, potentiellement issues d’un même événement d’isolement géographique. Dans le<br />

cas <strong>de</strong> la barrière équatoriale <strong>de</strong> l’EPR, les dates <strong>de</strong> divergence coïnci<strong>de</strong>nt relativement<br />

bien même s’il existe une forte variance associée due à la stochasticité du processus <strong>de</strong> tri<br />

<strong>de</strong>s allèles sur une pério<strong>de</strong> <strong>de</strong> temps aussi courte (1-2 millions d’années). Il est en effet<br />

évi<strong>de</strong>nt que plus l’événement <strong>de</strong> spéciation est ancien et plus la proportion <strong>de</strong> gènes ayant<br />

atteint la monophilie réciproque est élevée, ce qui fournit une meilleure estimation interespèces<br />

(Neigel & Avise 1997). Cette estimation peut être néanmoins rapi<strong>de</strong>ment biaisée si<br />

le temps <strong>de</strong> divergence <strong>de</strong>vient trop long et génère une saturation du signal phylogénétique<br />

au niveau du gène étudié (Avise 1994). L’approche comparative d’Hickerson et al (2006)<br />

avec le logiciel MsBayes, apparaît beaucoup plus informative. Sous l’hypothèse d’un taux<br />

103


Chapitre 3<br />

<strong>de</strong> mutation constant au cours du temps (horloge moléculaire), elle permet <strong>de</strong> tester le<br />

nombre d’évènements d’isolement génétique et <strong>de</strong> déterminer les dates les plus probables<br />

auxquelles ses évènements <strong>de</strong> vicariance ont eu lieu en imposant une histoire<br />

démographique propre à chaque lignée sœur.<br />

De plus, la phylogéographie comparée permet <strong>de</strong> préciser le positionnement <strong>de</strong> la<br />

barrière aux flux <strong>de</strong> gènes lorsque les espèces étudiées possè<strong>de</strong>nt <strong>de</strong>s capacités <strong>de</strong><br />

dispersion légèrement différentes (e.g. Maggs et al 2008, l’étu<strong>de</strong> d’une seule espèce<br />

pouvant engendrer une mauvaise interprétation <strong>de</strong> la position géographique historique <strong>de</strong><br />

cette barrière). L’approche comparative d’Hickerson et al (2006b) avec le logiciel<br />

MsBayes, est beaucoup plus informative. Elle permet <strong>de</strong> tester le nombre d’évènements<br />

d’isolement génétique et <strong>de</strong> déterminer les dates les plus probables auxquelles ses<br />

évènements <strong>de</strong> vicariance ont eu lieu.<br />

4.2. Apports <strong>de</strong> la phylogéographie comparée dans l’étu<strong>de</strong> <strong>de</strong> l’histoire<br />

démographique <strong>de</strong>s populations<br />

Outre les informations <strong>de</strong> vicariance, la phylogéographie comparée peut également<br />

s’avérer un outil puissant <strong>pour</strong> détecter <strong>de</strong>s évènements démographiques et notamment <strong>de</strong><br />

distinguer l’effet sélectif <strong>de</strong> l’effet démographique si ce <strong>de</strong>rnier a été causé par une<br />

perturbation environnementale « globale » ayant affectée l’ensemble <strong>de</strong> la faune (e.g.<br />

glaciations, éruption volcanique, etc…). En effet, la topologie d’un coalescent à un gène<br />

donné <strong>pour</strong> une espèce donnée ne permet pas <strong>de</strong> dissocier les causes ayant engendrées un<br />

écart à l’accumulation neutre <strong>de</strong>s mutations dans le polymorphisme. Si les tests d’écarts à<br />

la neutralité (tests <strong>de</strong> Tajima, Fu & Li ou Fay & Wu) sont congruents entre plusieurs<br />

espèces <strong>pour</strong> une zone géographique donnée ou une lignée intra-spécifique donnée ), il est<br />

alors peu probable que cet écart soit dû à <strong>de</strong>s évènements sélectifs concomitants sur le gène<br />

mais plus à une expansion démographique ou un goulot d’étranglement récent(e) affectant<br />

toutes les espèces (Zink 2002). Dans notre cas (espèces hydrothermales <strong>de</strong> l’EPR),<br />

l’hypothèse la plus parcimonieuse est alors une modification <strong>de</strong> l’habitat ayant engendré<br />

un événement démographique particulier sur l’ensemble <strong>de</strong>s espèces. Cette réduction <strong>de</strong> la<br />

taille <strong>de</strong>s populations (goulot d’étranglement ou bottleneck) suivie d’une expansion<br />

démographique à partir <strong>de</strong>s individus restants semble ici liée à une réduction <strong>de</strong> l’activité<br />

104


Chapitre 3<br />

hydrothermale ou à une série d’épanchements <strong>de</strong> magma (volcanisme effusif) ayant<br />

engendré une <strong>de</strong>struction importante <strong>de</strong> la faune associée. L’autre hypothèse d’une<br />

colonisation récente <strong>de</strong> l’EPR sud à partir d’une succession d’effets fondateurs du nord<br />

vers le sud est beaucoup plus difficile à considérer eu égard à la disparité <strong>de</strong>s mo<strong>de</strong>s <strong>de</strong><br />

dispersion observés chez les espèces étudiées et à l’absence d’une réelle migration<br />

asymétrique orientée du nord vers le sud au travers <strong>de</strong> la barrière équatoriale.<br />

4.3. Limites <strong>de</strong> la phylogéographie comparée monogène<br />

Si la comparaison <strong>de</strong> patrons phylogéographiques concordants entre espèces peut<br />

permettre <strong>de</strong> mieux évaluer la date <strong>de</strong> l’événement d’isolement, les étu<strong>de</strong>s intégratives,<br />

telles que celles d’Hickerson et al (2006), nécessitent <strong>de</strong> comparer le même gène <strong>pour</strong> un<br />

ensemble d’espèces ayant la même histoire et reposent sur un modèle <strong>de</strong> spéciation avec<br />

changements démographiques indépendants. Or, l’obtention <strong>de</strong> séquences à un même gène<br />

sur plusieurs espèces peut s’avérer techniquement difficile en absence d’amorces dites<br />

« universelles ». Pour cette raison, les approches intégratives <strong>de</strong> phylogéographie<br />

comparée ont, <strong>pour</strong> le moment, été menées presque exclusivement sur le gène<br />

mitochondrial <strong>de</strong> la Cytochrome Oxydase I (Hickerson et al 2006a, Hickerson & Meyer<br />

2008, Plouviez et al 2009 ; exception : Hickerson et al 2006b, étu<strong>de</strong> menée sur <strong>de</strong>s introns<br />

peu variables). Or, les gènes mitochondriaux présentent généralement une héritabilité<br />

uniquement maternelle (mais voir par exemple, Skibinski et al 1994 <strong>pour</strong> la double<br />

héritabilité mitochondrial). Ceci implique que les patrons phylogéographiques observés par<br />

cette approche intégrative ne reflètent que la distribution <strong>de</strong>s lignées femelles. L’utilisation<br />

<strong>de</strong> gènes nucléaires (hérédité biparentale) <strong>pour</strong>raient donc complémenter ces approches.<br />

De même, le développement <strong>de</strong> nouvelles métho<strong>de</strong>s d’analyses permettant d’inclure<br />

plusieurs gènes dans les approches intégratives sur plusieurs espèces, <strong>pour</strong>raient minimiser<br />

les risques d’observer <strong>de</strong>s patrons <strong>de</strong> distribution propres au gène étudié et non<br />

représentatif <strong>de</strong> l’espèce.<br />

105


Chapitre 3<br />

106


Chapitre 4<br />

Chapitre 4 : Étu<strong>de</strong> d’une barrière semi-permeable à la<br />

dispersion par une approche multi-locus sur plusieurs<br />

espèces<br />

1. Introduction à l’étu<strong>de</strong> d’une barrière à la dispersion par une approche<br />

multi-locus<br />

L’utilisation d’un seul locus dans les étu<strong>de</strong>s <strong>de</strong> phylogéographie pouvant conduire à<br />

<strong>de</strong>s conclusions erronées, en inférant l’histoire du gène plutôt que l’histoire <strong>de</strong> l’espèce<br />

(Hudson 1983, Pamilo & Nei 1988, Maddison 1997), Avise & Ball (1990) soulignent la<br />

nécessité <strong>de</strong> tester la concordance du signal phylogéographique entre plusieurs locus<br />

indépendants afin <strong>de</strong> reconstruire l’histoire évolutive <strong>de</strong>s populations. Ainsi, par exemple,<br />

si plusieurs locus regroupent <strong>de</strong> la même manière <strong>de</strong>s populations différenciées<br />

génétiquement <strong>de</strong> part et d’autre d’une barrière à la dispersion préalablement supposée,<br />

cette concordance entre locus apporte du crédit à l’existence d’une telle barrière et permet<br />

éventuellement <strong>de</strong> mieux situer sa position géographique. Néanmoins, la stochasticité du<br />

processus <strong>de</strong> coalescence implique que, <strong>pour</strong> que plusieurs gènes présentent une<br />

différenciation génétique concordante à travers une barrière (accumulation <strong>de</strong> divergence<br />

sur <strong>de</strong>s gènes différents), la cessation <strong>de</strong>s flux <strong>de</strong> gènes doit être relativement longue<br />

(Avise & Ball 1990). Il n’en <strong>de</strong>meure pas moins vrai qu’une convergence dans la disparité<br />

<strong>de</strong>s distributions alléliques (analyse <strong>de</strong>s fréquences alléliques) peut être également<br />

observée à une échelle <strong>de</strong> temps moins gran<strong>de</strong> et conduire aux mêmes conclusions si les<br />

effets dynamiques liés au fonctionnement <strong>de</strong>s populations sont négligeables (effet<br />

« métapopulation »).<br />

Le nombre optimal <strong>de</strong> locus <strong>pour</strong> estimer <strong>de</strong>s paramètres <strong>de</strong> taille efficace<br />

(Pluzhnikov & Donnelly 1996, Felsenstein 2006, Carling & Brumfield 2007), <strong>de</strong> flux <strong>de</strong><br />

gènes (Beerli & Felsenstein 1999) ou <strong>de</strong> taux d’expansion démographique (Kuhner et al<br />

1998), dépend <strong>de</strong> la complexité <strong>de</strong> la dynamique <strong>de</strong> ces populations, <strong>de</strong> l’écart potentiel<br />

<strong>de</strong>s gènes étudiés au modèle d’évolution neutre et, <strong>de</strong> l’histoire évolutive <strong>de</strong> l’espèce<br />

107


Chapitre 4<br />

considérée (Edwards & Beerli 2000, Hickerson et al 2007, Peters et al 2007). Felsenstein<br />

(2006) suggère néanmoins, qu’il est généralement plus informatif d’augmenter le nombre<br />

<strong>de</strong> locus indépendants considérés plutôt que le nombre d’individus ou la longueur <strong>de</strong>s<br />

séquences utilisées.<br />

Du fait <strong>de</strong> leur taille efficace plus faible liée à leur héritabilité généralement<br />

uniquement maternelle, les marqueurs cytoplasmiques (mitochondriaux mais également<br />

chloroplastiques chez les végétaux) ont une plus gran<strong>de</strong> probabilité <strong>de</strong> montrer une<br />

structure phylogéographique prononcée que les marqueurs nucléaires (Moore 1995). Ainsi,<br />

quand l’ADN mitochondrial présente une monophylie réciproque entre <strong>de</strong>ux populations,<br />

les gènes nucléaires peuvent n’arborer qu’une para- ou polyphylie (Fig. 4.1, Zink &<br />

Barrowclough 2008). L’accumulation et la fixation <strong>de</strong>s mutations dans le polymorphisme<br />

ne dépend pas du taux <strong>de</strong> mutation du gène considéré, mais presque exclusivement <strong>de</strong> sa<br />

taille efficace (Zink & Barrowclough 2008).<br />

108


Chapitre 4<br />

Fig. 4.1 Détectabilité <strong>de</strong>s évènements <strong>de</strong> vicariance et coalescents <strong>de</strong> gènes<br />

mitochondriaux et nucléaires associés sous l’hypothèse neutre. Adaptée <strong>de</strong> Zink &<br />

Barrowclough (2008).<br />

Les mutations sont représentées sur les coalescents, permettant <strong>de</strong> comprendre que le taux<br />

<strong>de</strong> mutations n’affecte pas le temps <strong>de</strong> coalescence.<br />

Cependant, la structure phylogéographique observée par <strong>de</strong>s marqueurs cytoplasmiques a<br />

également plus <strong>de</strong> chance d’avoir une origine purement stochastique, sous l’effet <strong>de</strong> la<br />

dérive génétique (les séquences échantillonnées dans les populations ont une chance plus<br />

gran<strong>de</strong> d’apparaître sous une monophylie réciproque car la monophylie réciproque est plus<br />

rapi<strong>de</strong>ment atteinte <strong>pour</strong> ce type <strong>de</strong> marqueur à taille efficace Ne, Kuo et Avise 2005,<br />

Irwin 2002, Zink & Barrowclough 2008) et nécessite donc d’être confirmée par <strong>de</strong>s locus<br />

nucléaires indépendants afin d’inférer l’histoire phylogéographique <strong>de</strong> l’espèce. De plus,<br />

les marqueurs à hérédité uniparentale ne tracent l’histoire que d’un sexe. Or, mâle et<br />

femelle peuvent parfois avoir subi <strong>de</strong>s histoires évolutives différentes (e.g. migration<br />

différentielle entre sexes, phylopatry, Hoelzer 1997) et ainsi extrapoler cette histoire aux<br />

<strong>de</strong>ux sexes peut engendrer <strong>de</strong>s conclusions erronées.<br />

Parce qu’ils reflètent à la fois l’évolution <strong>de</strong>s mâles et <strong>de</strong>s femelles, les gènes<br />

nucléaires peuvent donc améliorer la compréhension <strong>de</strong> l’histoire phylogéographique <strong>de</strong><br />

109


Chapitre 4<br />

l’espèce étudiée, mais également permettre d’évaluer la perméabilité d’une barrière<br />

génétique via l’étu<strong>de</strong> <strong>de</strong> l’hybridation potentielle entre lignées divergentes. Si un individu<br />

hybri<strong>de</strong> est capable <strong>de</strong> donner une <strong>de</strong>scendance (introgression), ces individus introgressés<br />

peuvent influencer l’histoire <strong>de</strong>s lignées <strong>de</strong> différentes manières (cf. revue <strong>de</strong> Seehausen<br />

2004) : (1) engendrer l’homogénéisation génétique <strong>de</strong> ces lignées, (2) le transfert <strong>de</strong><br />

matériel génétique entre lignées, potentiellement favorisé par <strong>de</strong> la sélection positive, (3)<br />

un renforcement <strong>de</strong> l’isolement reproducteur entre lignées. Cependant, ces individus<br />

peuvent également être retenus dans une zone <strong>de</strong> contact secondaire entre les lignées<br />

divergentes (zone d’hybridation) suivant un équilibre entre la dispersion et la sélection<br />

contre les hybri<strong>de</strong>s (Barton & Hewitt 1985). L’étendue géographique d’une zone<br />

d’hybridation dépend <strong>de</strong> cet équilibre et peut ainsi permettre <strong>de</strong> mieux comprendre<br />

l’impact d’une barrière génétique dans la structure <strong>de</strong>s populations.<br />

Dans le chapitre précé<strong>de</strong>nt, nous avons mis en évi<strong>de</strong>nce la présence d’un événement<br />

<strong>de</strong> vicariance partagé entre sept espèces (1,3 Ma) le long <strong>de</strong> la dorsale du Pacifique oriental<br />

sud laissant supposer l’existence d’une spéciation <strong>de</strong> type allopatrique en cours et, d’un<br />

second événement plus ancien (11,6 Ma) ayant conduit à une spéciation chez L. elevatus<br />

au niveau <strong>de</strong> la dorsale EPR nord (décalage géographique <strong>de</strong> la zone <strong>de</strong> contact<br />

secondaire). Néanmoins, l’utilisation du seul gène mtCOI ne nous a pas permis <strong>de</strong><br />

déterminer avec précision l’emplacement <strong>de</strong> la barrière aux flux <strong>de</strong> gènes <strong>pour</strong> chacune <strong>de</strong><br />

ces espèces. En effet, si A. pompejana et E. vitrea présentent une monophylie réciproque<br />

<strong>de</strong> part et d’autre <strong>de</strong> l’équateur, les lignées divergentes <strong>de</strong> B. thermophilus, par exemple, se<br />

distribuent selon un cline géographique le long <strong>de</strong> l’EPR, la lignée majoritairement<br />

distribuée au Nord n’étant pratiquement pas échantillonnée <strong>de</strong> 14°S à 21°33’S (2 individus<br />

à 17°25’S). La question <strong>de</strong> savoir si la présence <strong>de</strong> lignées mitochondriales divergentes<br />

traduit une spéciation en cours avec ou non la formation <strong>de</strong> zones <strong>de</strong> contact secondaire<br />

mérite également toute notre attention. De la même façon, sous l’hypothèse d’un contact<br />

secondaire (i.e. L. elevatus, L. ovalis ou B. thermophilus), il semble également important<br />

<strong>de</strong> déterminer si ces lignées sont capables <strong>de</strong> s’hybri<strong>de</strong>r, <strong>de</strong> préciser la nature <strong>de</strong><br />

l’hybridation et d’estimer le niveau d’introgression entre celles-ci. De plus, la détection<br />

préalable d’hybri<strong>de</strong>s via <strong>de</strong>s marqueurs allozymiques (Matabos et al 2008b) entre les<br />

espèces cryptiques <strong>de</strong> L. elevatus dans une zone <strong>de</strong> contact située à 9°50’N pose la<br />

question <strong>de</strong> l’étendue <strong>de</strong> cette zone d’hybridation (seules les populations <strong>de</strong> 13°N, 9°50’N<br />

et 17°25’S avaient été génotypées par ces auteurs). Dans ce présent chapitre, nous avons<br />

110


Chapitre 4<br />

donc choisi <strong>de</strong> comparer les patrons phylogéographiques contrastés révélés par le mtCOI à<br />

ceux obtenus sur <strong>de</strong>s marqueurs nucléaires chez A. pompejana, B. thermophilus et L.<br />

elevatus, appartenant à <strong>de</strong>s classes distinctes et <strong>pour</strong> lesquelles <strong>de</strong>s banques d’ADN<br />

complémentaires nous permettaient d’accé<strong>de</strong>r à <strong>de</strong> tels marqueurs (cf. Chapitre 2). Ces<br />

analyses phylogéographiques multi-locus seront présentées après une brève présentation <strong>de</strong><br />

la taxonomie et <strong>de</strong>s principaux traits d’histoire <strong>de</strong> vie <strong>de</strong>s espèces cibles.<br />

2. Taxonomie et traits d’histoire <strong>de</strong> vie <strong>de</strong>s espèces cibles<br />

La phase larvaire constitue le vecteur principal <strong>de</strong> colonisation <strong>de</strong>s espèces à cycle<br />

bentho-pélagique (Pechenik 1999). Néanmoins, la capacité natatoire <strong>de</strong>s larves étant<br />

extrêmement limitée, cette dispersion s’effectue essentiellement <strong>de</strong> manière passive par<br />

l’intermédiaire <strong>de</strong>s courants <strong>de</strong> fond orientés dans l’axe du graben (Chevaldonné et al 1997).<br />

Ainsi, toute modification topographique <strong>de</strong> la dorsale (segmentation <strong>de</strong> la vallée axiale par<br />

<strong>de</strong>s failles transformantes, microplaque) ou altération <strong>de</strong> la circulation océanique (inflations<br />

bathymétriques) peuvent modifier l’écoulement <strong>de</strong>s eaux <strong>de</strong> fond et, engendrer <strong>de</strong>s barrières<br />

à la dispersion. Ces barrières peuvent être responsables <strong>de</strong> véritables discontinuités<br />

génétiques, faisant du milieu hydrothermal un environnement propice à la spéciation par<br />

vicariance à l’échelle <strong>de</strong> la dorsale océanique.<br />

2.1. A. pompejana<br />

La famille <strong>de</strong>s Alvinellidae est endémique <strong>de</strong>s sources hydrothermales <strong>de</strong> l’océan<br />

Pacifique (Desbruyères & Laubier 1991) et se compose d’une douzaine d’espèces dont<br />

<strong>de</strong>ux appartenant au genre Alvinella (Desbruyères et al 2006b) : A. pompejana et A.<br />

caudata (les autres étant du genre Paralvinella), trouvés uniquement sur l’EPR. La<br />

séparation phylogénétique <strong>de</strong>s espèces sur la base <strong>de</strong> leurs isoformes enzymatiques a<br />

montré que les regroupements d’espèces ne se faisaient pas en fonction <strong>de</strong> leur proximité<br />

géographique mais en plutôt en fonction <strong>de</strong> l’habitat (notamment le preferendum<br />

thermique : Jollivet et al 1995) laissant supposer qu’une spéciation <strong>de</strong> type ‘écologique’ ait<br />

pu prédater une spéciation <strong>de</strong> type allopatrique (existence d’espèces jumelles entre EPR et<br />

dorsale nord Pacifique : Tunnicliffe 1991).<br />

111


Chapitre 4<br />

Le polychète tubicole Alvinella pompejana a été observé à <strong>de</strong>s températures pouvant<br />

atteindre 105°C sur <strong>de</strong> courtes pério<strong>de</strong>s (enroulement d’un individu sur une son<strong>de</strong><br />

indiquant 105°C), avec <strong>de</strong>s températures à l’ouverture du tube <strong>de</strong> 20-45°C (Chevaldonné et<br />

al 2002). Cette espèce a ainsi été décrite comme l’un <strong>de</strong>s métazoaires les plus tolérants aux<br />

températures élevées. A. pompejana est une espèce pionnière dans la colonisation <strong>de</strong><br />

nouvelles sources hydrothermales (Jollivet 1993, Desbruyères et al 1998). Cette espèce<br />

présente <strong>de</strong>s oocytes <strong>de</strong> 200 µm riche en vitellus. La fécondation <strong>de</strong>s ovocytes est interne<br />

et fait suite à un appariement <strong>de</strong>s mâles et <strong>de</strong>s femelles au cours duquel le sperme du mâle<br />

est transféré à la femelle tête à tête. Les spermatozoï<strong>de</strong>s sont transférés dans une<br />

spermathèque et fécon<strong>de</strong>nt les ovocytes matures au moment <strong>de</strong> leur émission dans la<br />

colonne d’eau <strong>pour</strong> donner lieu à <strong>de</strong>s larves lécithotrophes (Chevaldonné & Jollivet 1993,<br />

Chevaldonné et al 1996, Jollivet 1996, Pradillon & Gaill 2003, Fig. 4.2). Le ralentissement<br />

du développement embryonnaire en eau froi<strong>de</strong>, détecté expérimentalement chez cette<br />

espèce, retar<strong>de</strong>rait la métamorphose <strong>de</strong>s larves et <strong>pour</strong>rait ainsi, en augmentant la durée <strong>de</strong><br />

dispersion dans la colonne d’eau, permettre une dispersion sur <strong>de</strong> plus longues distances<br />

(dépendant alors <strong>de</strong>s courants profonds circulant à l’axe <strong>de</strong> la dorsale : Pradillon et al<br />

2001).<br />

112


Chapitre 4<br />

Adultes gonochoriques 1<br />

Sexe ratio équilibré 2 (matures<br />

biaisé en faveur <strong>de</strong>s mâles 3 )<br />

Femelle<br />

Oogenèse rapi<strong>de</strong><br />

continue 3 Vitellogenèse lente 3<br />

Mâle<br />

Oogonies<br />

prévitellogènes<br />

Juvénile<br />

Spermatogenèse<br />

sperme<br />

Oocytes avec vitellus<br />

(~200 µm) 4<br />

Recrutement<br />

discontinu 3<br />

Fécondation quasicontinue<br />

3<br />

Larves lécithotrophes<br />

dans la colonne d’eau 4<br />

Fig. 4.2 Cycle <strong>de</strong> vie d’A. pompejana.<br />

La longueur <strong>de</strong>s flèches n’est pas proportionnelle au temps.<br />

1 Zal et al (1994), 2 Pradillon et al (2005), 3 Faure et al (2007), 4 Chevaldonné et al<br />

(1996)<br />

2.2. B. thermophilus<br />

Les bivalves du genre Bathymodiolus constituent l’un <strong>de</strong>s taxons les plus<br />

cosmopolites et présentant la plus forte biomasse <strong>de</strong>s sources hydrothermales (importantes<br />

moulières recensées en <strong>de</strong> nombreux points du globe sur quasiment tous les sites<br />

hydrothermaux connus exception faite <strong>de</strong> la dorsale Nord Pacifique (Desbruyère et al<br />

2006b). Composés d’une vingtaine d’espèces (Won et al 2008), ce genre présente<br />

également une forte diversité spécifique et est également distribué au niveau <strong>de</strong>s zones <strong>de</strong><br />

113


Chapitre 4<br />

suintements froids. Les espèces du genre Bathymodiolus apparaissent clairement<br />

polyphylétiques et se trouvent mélangées à d’autres genres provenant d’habitats variés<br />

(carcasses <strong>de</strong> baleine, suintements et bois coulés) : Gigantidas, Idas, Adipicola (Samadi et<br />

al 2007, Lorion et al 2009). La branche évolutive dans laquelle se trouve B. thermophilus<br />

est, quant à elle, composée uniquement d’espèces appartenant au genre Bathymodiolus et<br />

<strong>pour</strong>rait être originaire d’un ancêtre commun ayant vécu dans les zones <strong>de</strong> suintements<br />

froids du Golfe du Mexique (Faure 2008). Cette espèce peut être considérée comme une<br />

espèce-architecte et abritent <strong>de</strong> nombreuses espèces <strong>de</strong> macrofaune <strong>de</strong> petite taille (e.g.<br />

gastéropo<strong>de</strong>s, polychètes). (Jollivet 1996).<br />

Le cycle <strong>de</strong> vie <strong>de</strong>s Bathymodiolus est présenté en figure 4.3. Les Bathymodiolus<br />

sont généralement gonochoriques ou hermaphrodites protandres, (e.g. B. elongatus, Le<br />

Pennec & Benniger 1997, B. azoricus Comtet et al 1999). La fécondation <strong>de</strong>s gamètes<br />

s’effectue dans la colonne d’eau <strong>pour</strong> former <strong>de</strong>s œufs (50 µm) puis <strong>de</strong>s larves<br />

planctotrophes (Lutz et al 1980, Le Pennec 1988). Ces larves sont capables <strong>de</strong> remonter<br />

dans la colonne d’eau jusqu’à atteindre les couches éclairées <strong>de</strong> l’océan (Arellano &<br />

Young 2009), plus riches en nourriture. Il a été ainsi montrer que les larves pouvaient<br />

passer environ 4 mois dans la colonne d’eau entre leur émission (prodissochonque I : 80<br />

µm) et leur recrutement (prodissochonque II : 400 µm) (Dixon et al 2007). Le recrutement<br />

a été suggéré discontinu chez B. azoricus par Comtet & Desbruyères (1998) mais continu<br />

chez B. thermophilus par Van Dover et al (2001b). Par cette remontée verticale, les larves<br />

<strong>pour</strong>ront alors être soumises à <strong>de</strong>s courants océaniques différents <strong>de</strong>s courants profonds,<br />

par leur vitesse, mais également leur direction (Thomson et al 1990, Cannon & Pashinski<br />

1997). L’accès à la dispersion via ces courants moins profonds <strong>pour</strong>rait ainsi favoriser la<br />

migration à travers une barrière physique et permettre un contact secondaire entre lignées<br />

divergentes.<br />

114


Chapitre 4<br />

Adultes généralement<br />

gonochoriques<br />

ou<br />

Hermaphrodisme protandre 1<br />

Adulte<br />

ou<br />

mature<br />

Juvénile<br />

(~400 µm) 6<br />

Gamétogenèse<br />

Oogonies<br />

Sperme<br />

discontinue 1 Fécondation externe 2,3<br />

Recrutement<br />

Larves prodissochonque II<br />

(400 µm) 4<br />

Migration verticale dans<br />

la colonne d’eau 5<br />

Oocytes (50 µm) 2,3<br />

Larves planctotrophe<br />

(prodissochonque I : 80 µm) 4<br />

Émission <strong>de</strong>s larves<br />

Fig. 4.3 Cycle <strong>de</strong> vie <strong>de</strong> B. thermophilus<br />

La longueur <strong>de</strong>s flèches n’est pas proportionnelle au temps.<br />

1 Le Pennec & Beninger (1997), 2 Lutz et al (1980), 3 Le Pennec (1988), 4 Dixon et<br />

al (2007), 5 Arellano & Young (2009), 6 Mullineaux et al (1998)<br />

115


Chapitre 4<br />

2.3. L. elevatus<br />

La famille <strong>de</strong>s Lepetodrilidae représente une <strong>de</strong>s familles <strong>de</strong> gastéropo<strong>de</strong>s<br />

patelliformes les plus abondantes <strong>de</strong>s sources hydrothermales profon<strong>de</strong>s. Certaines espèces<br />

ont également été décrites sur les bois coulés (Lepetodrilus elevatus McLean 22°N, EPR)<br />

et carcasses <strong>de</strong> baleines (L. pustulosus McLean, Monterey Bay, 36°N, Johnson et al 2008).<br />

De plus, cette famille compte également une forte diversité spécifique avec actuellement<br />

plus <strong>de</strong> 20 espèces morphologiquement distinctes (Warén & Boucher 2001, Desbruyères et<br />

al 2006b) dont la plupart appartiennent au genre Lepetodrilus. Venant s’ajouter à ce<br />

décompte, <strong>de</strong> nombreux complexes d’espèces (e.g. complexe « fucensis » : 2 espèces,<br />

Johnson et al 2006 ; complexe « elevatus » : 4 espèces, Matabos et al 2008b, Johnson et al<br />

2008 ; complexe « schrolli » : 3 espèces, Johnson et al 2008) ont été mis en évi<strong>de</strong>nce au<br />

sein <strong>de</strong> ce genre. Originaire du Mésozoïque (McLean 1988), les Lepetodrilidae, et plus<br />

particulièrement le genre Lepetodrilus, ont donc connu <strong>de</strong>puis lors une véritable radiation<br />

évolutive, suscitant un grand intérêt <strong>pour</strong> l’étu<strong>de</strong> <strong>de</strong>s processus <strong>de</strong> spéciation ayant conduit<br />

à une telle diversification.<br />

Les Lepetodrilus adultes se déplacent relativement lentement (< 0,1 cm/s, observation<br />

personnelle sur boîte <strong>de</strong> pétri sur L. schrolli) et ont une activité plutôt grégaire et sé<strong>de</strong>ntaire<br />

(empreintes <strong>de</strong>s coquilles sur les tubes <strong>de</strong> vestimentifères dues au broutage <strong>de</strong>s bactéries<br />

sur le tube (cf. L. elevatus et L. pustulosus). Leur transport via la faune vagile (phorésie),<br />

notamment par les crabes hydrothermaux (cf. Johnson et al 2008), constitue un vecteur <strong>de</strong><br />

dispersion efficace mais limité à l’échelle du champ hydrothermal (dispersion <strong>de</strong> proche en<br />

proche souvent liée à la migration massive <strong>de</strong>s crabes lors <strong>de</strong> l’extinction d’un site<br />

d’activité : Jollivet 1993). Au contraire, l’aptitu<strong>de</strong> <strong>de</strong>s larves à entrer en dormance ou à<br />

présenter d’importantes réserves vitellines leur permet <strong>de</strong> pallier au manque <strong>de</strong> nourriture<br />

dans l’eau <strong>de</strong> fond oligotrophe et donc <strong>de</strong> disperser sur <strong>de</strong> longues distances (Jollivet et al<br />

2000 ; Pradillon et al 2001 ; Metaxas et al 2004). Le cycle <strong>de</strong> vie <strong>de</strong> L. elevatus est présenté<br />

en figure 4.4.<br />

116


Chapitre 4<br />

Adultes gonochoriques<br />

Sexe ratio équilibrée 7<br />

Femelle<br />

Mâle<br />

Gamétogenèse<br />

continue 1<br />

Oogonies<br />

prévitellogènes<br />

Juvénile<br />

(~560 µm) 5<br />

Recrutement<br />

discontinu 6<br />

Sperme<br />

Vitellogenèse<br />

continue 2<br />

Oocytes avec vitellus<br />

(max 90 µm,<br />

moyenne 56 µm) 2<br />

Fécondation<br />

« interne » continue 3<br />

Larves (~170 µm) dans la<br />

plume hydrothermale 4, 5<br />

Larves lécithotrophes<br />

incubées dans la cavité<br />

palléale <strong>de</strong>s femelles 3<br />

Libération <strong>de</strong>s larves<br />

Fig. 4.4 Cycle <strong>de</strong> vie <strong>de</strong> L. elevatus.<br />

La longueur <strong>de</strong>s flèches n’est pas proportionnelle au temps.<br />

1 Kelly & Metaxas (2007), 2 Tyler et al (2008), 3 Fretter (1988), 4 Mullineaux et al<br />

(1995), 5 Mullineaux et al (1996), 6 Mullineaux et al (1998), 7 Matabos et al (2008a).<br />

Une espèce <strong>de</strong> Lepetodrilus donnée est généralement présente sur une seule dorsale ou<br />

un segment <strong>de</strong> dorsale et un habitat particulier, pouvant suggérer une diversification liée à<br />

<strong>de</strong>s processus <strong>de</strong> vicariance. Néanmoins, certaines espèces présentent <strong>de</strong>s aires <strong>de</strong><br />

distribution superposées (e.g. L. elevatus, L. pustulosus, L. ovalis, L. cristatus sur l’EPR)<br />

au sein d’un segment <strong>de</strong> dorsale mais ayant un habitat clairement défini (tubes <strong>de</strong><br />

vestimentifères, cheminée, basaltes ou coquilles <strong>de</strong> bivales). Dans certains cas, les espèces<br />

se répartissent différemment à l’échelle d’un site, selon leur preferendum thermique<br />

(Bates et al 2005, Mills et al 2007) mais aussi potentiellement selon la composition<br />

117


Chapitre 4<br />

physico-chimique du flui<strong>de</strong> et/ou <strong>de</strong>s relations interspécifiques entre compétiteurs proches<br />

(Matabos et al 2008a). Ces différences <strong>de</strong> distribution liées à l’environnement local<br />

<strong>pour</strong>raient ainsi favoriser la spéciation écologique (ou parapatrique).<br />

3. Matériels et métho<strong>de</strong>s <strong>de</strong>s espèces cibles<br />

L’extraction d’ADN, l’amplification et le séquençage du gène mtCOI ont été réalisés<br />

selon le protocole décrit dans Plouviez et al (2009). Les gènes nucléaires <strong>de</strong>s différentes<br />

espèces cibles ont été amplifiés selon les conditions décrites dans l’annexe 3. Le clonage<br />

<strong>de</strong> ces gènes nucléaires a été réalisé par Marquage-Recapture <strong>de</strong> séquences (Bierne et al<br />

2007 présenté en annexe 1 et chapitre 2) selon le même protocole <strong>de</strong> clonage <strong>pour</strong> A.<br />

pompejana, B. thermophilus et L. elevatus. Généralement, <strong>de</strong>ux clonages comprenant<br />

chacun 8 individus ont été réalisés <strong>pour</strong> chaque population et 32 clones par clonage ont été<br />

séquencés. Concernant B. thermophilus, les séquences <strong>de</strong>s individus échantillonnés à<br />

7°25’S et 21°33’S ont été réalisées par Dr Baptiste Faure sur un clonage unique par<br />

population <strong>de</strong> 48 individus avec un effort <strong>de</strong> recapture <strong>de</strong> X1 (i.e. 96 clones séquencés par<br />

clonage). Pour L. elevatus, le nombre d’individus clonés a été augmenté dans la zone <strong>de</strong><br />

contact (9°50’N), ainsi qu’à 21° 33’S (où une troisième espèce cryptique <strong>de</strong> L. elevatus a<br />

été observée sur la base d’une divergence mitochondriale) : soient 32 individus à 9°50’N et<br />

24 à 21°33’S. L’effort <strong>de</strong> capture, quel que soit le nombre d’individus, est <strong>de</strong> X2 (e.g. 32<br />

séquences par clonage comprenant 8 individus). Les séquences ont ensuite été traitées<br />

selon le protocole décrit dans la partie ‘Matériels et Métho<strong>de</strong>s’ du chapitre 2 afin<br />

d’éliminer la recombinaison in vitro et <strong>de</strong> ne gar<strong>de</strong>r qu’un gène supposé unique et diploï<strong>de</strong>.<br />

4. Analyses et résultats sur les espèces cibles<br />

4.1. A.pompejana<br />

4.1.1. Résumé <strong>de</strong>s résultats<br />

Malgré une microstructure génétique significative <strong>de</strong> type chaotique au sein d’un<br />

champ hydrothermal donné (i.e. 13°N/EPR), <strong>de</strong>s étu<strong>de</strong>s allozymiques menées sur le<br />

polychète hydrothermal A. pompejana au nord <strong>de</strong> l’EPR (<strong>de</strong> 21°N à 9°50’N) n’ont pas<br />

118


Chapitre 4<br />

permis <strong>de</strong> détecter <strong>de</strong> réelle structure géographique entre ces champs hydrothermaux<br />

(Jollivet et al 1995, Jollivet 1996), suggérant <strong>de</strong> fortes capacités <strong>de</strong> dispersion chez cette<br />

espèce ou l’action <strong>de</strong> processus sélectifs homogénéisant. Cependant, un modèle fluxpropagule<br />

basé sur la biologie <strong>de</strong> cette espèce suggère que, si les larves migrent dans la<br />

colonne d’eau durant 15 à 30 jours, la renverse périodique <strong>de</strong>s courants <strong>de</strong> marée ne<br />

permet pas une migration efficace inter-champs <strong>de</strong> celles-ci (Chevaldonné et al 1997),<br />

impliquant une différenciation génétique <strong>de</strong>s populations entre ces champs par simple<br />

dérive. L’absence <strong>de</strong> structure génétique à l’échelle <strong>de</strong> l’EPR Nord <strong>pour</strong>rait alors<br />

s’expliquer par<br />

(1) un manque <strong>de</strong> résolution <strong>de</strong>s marqueurs enzymatiques,<br />

(2) une sélection balancée sur ces marqueurs maintenant une homogénéité <strong>de</strong>s<br />

fréquences allozymiques directement liée à la nature <strong>de</strong> l’habitat<br />

hydrothermal, comme le proposent Piccino et al (2004) <strong>pour</strong> le locus<br />

phosphoglucomutase (PGM), celui-ci présentant un allèle 90 thermorésistant<br />

plus avantageux dans les populations nouvellement fondées ou les conditions<br />

thermiques sont plus élevées mais en équi-fréquence avec l’allèle 100 dans les<br />

populations plus « agées »,<br />

(3) un déplacement <strong>de</strong> l’activité hydrothermale le long <strong>de</strong> l’axe (10 mètres par<br />

année : Watremez et al 1990, Kervévan 1991) pouvant, <strong>de</strong> manière transitoire,<br />

permettre la connectivité <strong>de</strong>s champs hydrothermaux par rapprochement et<br />

ainsi empêcher la différenciation génétique <strong>de</strong> s’établir dans le temps<br />

(modélisé par Jollivet et al 1999).<br />

Le séquençage du gène mtCOI sur <strong>de</strong>s individus issus <strong>de</strong> populations réparties sur<br />

l’ensemble <strong>de</strong> la dorsale EPR a permis <strong>de</strong> révéler la présence d’une barrière aux flux <strong>de</strong><br />

gènes à l’Equateur (Hurtado et al 2004, Plouviez et al 2009), donnant lieu à <strong>de</strong>ux cla<strong>de</strong>s<br />

divergents quasi-monophylétiquement réciproques entre les individus du Nord et du Sud<br />

<strong>de</strong> l’EPR. Nous avons donc cherché à déterminer l’impact <strong>de</strong> cette barrière datée à 1,3 Ma<br />

sur les lignées mitochondriales divergentes en utilisant à la fois les variations en fréquence<br />

<strong>de</strong>s marqueurs enzymatiques précé<strong>de</strong>mment utilisés uniquement dans la partie Nord <strong>de</strong><br />

l’EPR et l’histoire évolutive <strong>de</strong>s allèles (séquences) <strong>de</strong> trois gènes nucléaires (dont le gène<br />

PGM <strong>pour</strong> lequel une sélection <strong>de</strong>s formes 90 et 100 est suspectée).<br />

119


Chapitre 4<br />

Si les valeurs <strong>de</strong> différenciation génétique obtenues sur les marqueurs<br />

allozymiques sont proches <strong>de</strong> celles obtenues à micro-échelle (au sein d’un champ<br />

hydrothermal : 13°N/EPR), six <strong>de</strong> ces marqueurs permettent néanmoins <strong>de</strong> confirmer la<br />

présence d’une barrière aux flux <strong>de</strong> gènes à l’Equateur.<br />

D’une manière plus frappante, l’analyse <strong>de</strong>s parentés alléliques par <strong>de</strong>s approches <strong>de</strong><br />

coalescence (Fig. 4.5) montre <strong>de</strong> manière très claire l’existence d’une divergence entre les<br />

lignées alléliques du gène <strong>de</strong> la PGM (réseau d’allèles révélant 4 cla<strong>de</strong>s divergents : <strong>de</strong>ux<br />

échantillonnés majoritairement au Nord et les <strong>de</strong>ux autres échantillonnés majoritairement<br />

au Sud) et, à une échelle moindre entre 2 lignées d’allèles du gène codant <strong>pour</strong> la<br />

GlobineX (2 lignées d’allèles plus faiblement divergentes présentant <strong>de</strong>s fréquences<br />

alléliques différentes dans les <strong>de</strong>ux régions géographiques). Pour ce <strong>de</strong>rnier gène, ces<br />

différences <strong>de</strong> fréquences se traduisent par une différenciation génétique significative entre<br />

les populations <strong>de</strong> part et d’autre <strong>de</strong> cette barrière. À l’inverse, aucune divergence n’a été<br />

mise en évi<strong>de</strong>nce dans la phylogénie d’allèles associée au gène S-A<strong>de</strong>nosyl Homocystéine<br />

Hydrolase (SAHH) et aucune différenciation génétique Nord/Sud n’a été détectée sur la<br />

base <strong>de</strong>s fréquences alléliques observées.<br />

120


Chapitre 4<br />

0,0005 0,0005<br />

PGM<br />

0,001 0,001<br />

0,001<br />

GlobX<br />

SAHH<br />

Fig. 4.5 Arbres phylogénétiques réalisés selon la métho<strong>de</strong>s <strong>de</strong>s plus proches voisins sur<br />

les gènes nucléaires d’A. pompejana à partir <strong>de</strong>s jeux <strong>de</strong> séquences <strong>de</strong> qualité 2 (2<br />

allèles maximum par individu).<br />

En bleu et rouge : séquences d’individus échantillonnées respectivement au Nord et au Sud<br />

<strong>de</strong> l’EPR.<br />

121


Chapitre 4<br />

L’étu<strong>de</strong> <strong>de</strong> la migration entre ces <strong>de</strong>ux régions géographiques via le logiciel Migrate-n<br />

(Beerli & Felsenstein 2001) a révélé l’existence d’une migration asymétrique à travers<br />

l’Equateur, du Nord vers le Sud <strong>de</strong> la dorsale. Ce sens <strong>de</strong> migration <strong>pour</strong>rait expliquer la<br />

différenciation génétique observée <strong>de</strong>s champs hydrothermaux 7°25’S-14°S versus<br />

17°25’S-21°33’S sur les allozymes, détectée via le logiciel Structure (mise en évi<strong>de</strong>nce <strong>de</strong><br />

3 groupes génétiquement différenciés : Prichard et al 2000), le groupe 2 composé<br />

essentiellement d’individus <strong>de</strong> 7°25’S et 14°S <strong>pour</strong>rait indiqué que ces latitu<strong>de</strong>s<br />

correspon<strong>de</strong>nt à une zone <strong>de</strong> transition entre les populations du Nord <strong>de</strong> la dorsale<br />

(majoritaire dans le groupe 1) et les populations du Sud <strong>de</strong> la dorsale (majoritaire dans le<br />

groupe 3). L’existence <strong>de</strong> Fis significatifs au niveau <strong>de</strong>s locus enzymatiques sur les sites<br />

7°25S, 14°S et 17°25S renforce cette hypothèse <strong>de</strong> zone <strong>de</strong> contact secondaire. De plus,<br />

une analyse RFLP/typage d’un in<strong>de</strong>l basée sur les sites fixés <strong>de</strong>s 4 lignées alléliques du<br />

gène PGM ont permis <strong>de</strong> détecter <strong>de</strong>s individus hybri<strong>de</strong>s/introgressés majoritairement aux<br />

sites 7°25’S à 14°S, indiquant que ces sites <strong>pour</strong>raient correspondre à une zone<br />

d’hybridation entre les lignées mitochondriales.<br />

L’utilisation d’un modèle d’ « Isolement avec Migration » <strong>de</strong> Hey & Nielsen (2007) a<br />

révélé la présence d’une introgression asymétrique relativement faible entre les 2 lignées<br />

mitochondriales <strong>de</strong> la lignée Sud vers la lignée Nord avec un temps <strong>de</strong> divergence moyen<br />

estimé sur les gènes mtCOI et nucléaires qui concor<strong>de</strong> avec celui proposé par l’approche<br />

multispécifique au gène mtCOI (1,3 Ma, Plouviez et al 2009, Fig. 4.6).<br />

122


Chapitre 4<br />

A<br />

B<br />

C<br />

Fig. 4.6 Distributions <strong>de</strong>s probabilités postérieures marginales du temps <strong>de</strong><br />

divergence entre lignées (A), <strong>de</strong> la taille efficace <strong>de</strong> la population ancestrale et <strong>de</strong> ces<br />

populations filles, (B), et <strong>de</strong>s taux <strong>de</strong> flux <strong>de</strong> gènes directionnels entre lignées (C),<br />

estimées par un modèle d’Isolement avec Migration via IMa.<br />

123


Chapitre 4<br />

Enfin, les D <strong>de</strong> Tajima et les F* <strong>de</strong> Fi & Li, significatifs au Sud <strong>de</strong> la dorsale <strong>pour</strong><br />

l’ensemble <strong>de</strong>s gènes, ainsi qu’une analyse visant à discriminer l’effet d’une expansion<br />

démographique d’un balayage sélectif, ont permis <strong>de</strong> confirmer que l’accumulation non<br />

neutre <strong>de</strong>s mutations dans le polymorphisme <strong>de</strong> la lignée sud est bien due à une expansion<br />

démographique telle que précé<strong>de</strong>mment suggérée par Plouviez et al (2009). Cette <strong>de</strong>rnière<br />

analyse a également mis en évi<strong>de</strong>nce un balayage sélectif sur le gène SAHH, pouvant<br />

expliquer l’absence <strong>de</strong> différenciation génétique entre les populations Nord et le Sud <strong>de</strong><br />

l’EPR lorsque l’on considère ce gène uniquement. Ce <strong>de</strong>rnier résultat vali<strong>de</strong> l’hypothèse du<br />

passage facilité d’un allèle avantageux à travers une barrière physique à la dispersion<br />

lorsque les espèces ont encore peu divergées : une situation assez similaire à celle trouvée<br />

par Faure et al (2007) entre Mytilus edulis et M. galloprovincialis. Bien que n’étant pas<br />

détecté comme tel par le logiciel Sweep_bott (Galtier et al 2000), la topologie du<br />

coalescent du gène GlobX <strong>pour</strong>rait également être apparentée à un balayage sélectif plus<br />

récent (et donc moins détectable) sur ce gène.<br />

La comparaison <strong>de</strong>s génotypes RFLP <strong>de</strong>s 4 allèles du gène PGM avec les génotypes<br />

allozymiques a permis <strong>de</strong> corréler les cla<strong>de</strong>s 3 et 4 aux allèles 90 et 100 trouvés au Sud <strong>de</strong><br />

l’EPR et ainsi <strong>de</strong> montrer qu’une mutation Q/E non-synonyme engendrant une différence<br />

<strong>de</strong> charge nette entre les protéines sur l’exon 4 expliquait la différence <strong>de</strong> mobilité<br />

électrophorétique observée entre les allèles 90 et 100 au Sud. Cependant, cette corrélation<br />

n’étant pas trouvée au Nord, il semble que l’allélisme 90/100 observé dans les populations<br />

nord soit associé à une mutation affectant un autre site exonique du gène. Cette homoplasie<br />

<strong>de</strong> charge souligne ainsi une limite <strong>de</strong>s locus enzymatiques dans la détection <strong>de</strong> structure<br />

génétique et l’intérêt <strong>de</strong> coupler ces marqueurs à d’autres types <strong>de</strong> marqueurs, en<br />

particulier basés sur le polymorphisme <strong>de</strong> séquences. L’approche multilocus effectuée a<br />

ainsi permis d’apporter du crédit aux hypothèses d’un effet vicariant allopatrique et d’une<br />

expansion démographique dans la lignée Sud proposées par l’approche multispécifique au<br />

gène mtCOI, mais également <strong>de</strong> révéler le caractère semi-perméable, au moins<br />

transitoirement, <strong>de</strong> cette barrière à l’Equateur <strong>pour</strong> A. pompejana, malgré la monophylie<br />

réciproque observée sur le mtCOI.<br />

124


Chapitre 4<br />

4.1.2. Article en préparation<br />

Across the equatorial East Pacific Rise barrier: do mitochondrial lineages of the<br />

dwelling polychaete Alvinella pompejana represent true ongoing allopatric species?<br />

Plouviez S. 1,2 , Le Guen D. 1,2 , Lecompte, O. 3 , Lallier F.H. 1,2 , Jollivet D. 1,2<br />

1 Université Pierre et Marie Curie-Paris 6, Laboratoire Adaptation et Diversité en Milieu<br />

Marin, <strong>Roscoff</strong>, France<br />

2 CNRS, UMR 7144 <strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>, BP 74, Place Georges Teissier, 29682,<br />

<strong>Roscoff</strong>, France<br />

3 IGBMC, Laboratoire <strong>de</strong> Bioinformatique et <strong>de</strong> Génomique Intégratives, Parc <strong>de</strong><br />

l’Innovation, Campus Illkirsh, Strasbourg, France<br />

Key words: migration, geographic barrier, selective sweep, balancing selection, subdivi<strong>de</strong>d<br />

populations, <strong>de</strong>mographic expansion<br />

125


Chapitre 4<br />

Abstract<br />

Relaxation of the strength of a barrier to gene flow can lead to secondary contact zone<br />

between previously isolated populations. Providing that isolation was not too long to avoid<br />

reproductive isolation between genetically divergent lineages, hybridization and<br />

subsequent allele introgression can then occur between these lineages. Spread of alleles<br />

across the barrier would thus mainly <strong>de</strong>pend on selection against hybrids, which can affect<br />

genes differentially according to their linkage to isolation genes or the strength of natural<br />

selection: genes un<strong>de</strong>r selective sweep or balancing selection would tend to cross the<br />

barrier more rapidly than neutral ones whereas genes un<strong>de</strong>r strong bidirectional selection<br />

would take longer time to cross the barrier. Previous studies in the <strong>de</strong>ep-sea hydrothermal<br />

vent polychaete Alvinella pompejana showed the occurrence of a clear barrier to gene flow<br />

at the Equator along the East Pacific Rise (EPR) for the mitochondrial Cytochrome<br />

Oxidase I gene (mtCOI). In the present study, allozymes and two nuclear loci confirmed<br />

this strong North/South differentiation of populations along the EPR. In addition, a North<br />

to South migration of propagules leading to a South to North allele introgression pattern<br />

between the two main divergent mtCOI lineages was <strong>de</strong>tected and allowed us to i<strong>de</strong>ntify<br />

genes un<strong>de</strong>r various strength of selection. The spread of alleles and associated levels of<br />

introgression on a gene previously hypothesized to be un<strong>de</strong>r balancing selection<br />

(phosphoglucomutase) was also investigated. The multilocus analysis also confirmed the<br />

<strong>de</strong>mographic expansion in the South previously suggested by a multispecies<br />

phylogeographic approach.<br />

126


Chapitre 4<br />

Introduction<br />

Dispersal is one of the major evolutionary forces that counter-balances the<br />

geographical structuring of populations when looking at the species range. Many<br />

population mo<strong>de</strong>ls have been proposed to <strong>de</strong>scribe how species disperse between<br />

geographically separated localities from the n-island mo<strong>de</strong>l (Wright 1931, 1937) to the<br />

stepping-stone (Kimura 1953) or the isolation-by-distance (Wright 1943, Malécot 1950)<br />

mo<strong>de</strong>ls. These mo<strong>de</strong>ls were then adapted into a metapopulation context to integrate<br />

population dynamics through time (i.e. local extinction and recolonisation rates, reviewed<br />

in Harrison & Hastings 1996). Un<strong>de</strong>r a ‘migrant’ metapopulation mo<strong>de</strong>l (Slatkin 1977,<br />

colonisation of a new patch comes from migrants originated from the whole<br />

metapopulation), local populations tend to be genetically homogenised over the whole<br />

metapopulation, providing that the number of colonists is twice greater as ‘true’ migrants.<br />

However, un<strong>de</strong>r ‘propagule’ metapopulation mo<strong>de</strong>l (Slatkin 1977, colonisation of a new<br />

patch comes from migrants of neighbouring sites), local extinction always increases<br />

genetic differentiation between <strong>de</strong>mes even for high levels of gene flow (see Whitlock &<br />

McCauley 1990, Pannel & Charlesworth 2000). The relationship between genetic<br />

differentiation and the migration rate is however seriously affected by the raise of a<br />

physical barrier to dispersal. Whatever the mo<strong>de</strong>l used, the setting up of a physical barrier<br />

to dispersal over a long period of time abruptly reduces or stops gene flow between the<br />

isolated sites and leads to a rapid change in allele frequencies between each part of the<br />

isolation point. Such a genetic differentiation is often associated with the fixation of<br />

mutations (divergence) between two sets of alleles (cla<strong>de</strong>s) leading to reciprocal allele<br />

monophylies for most genes un<strong>de</strong>r scrutiny providing that no lineage extinction or<br />

population admixture occurs secondarily (Cunningham & Collins 1998). In case of<br />

secondary contact, a genetic barrier may be superimposed onto this dispersal break<br />

possibly leading to a clinal distribution of alleles whose shape mainly <strong>de</strong>pends on the<br />

strength of the selection against hybrids (Barton & Hewitt 1985).<br />

Moreover, allele exchanges across a barrier may be dampered or accelerated <strong>de</strong>pending<br />

on how selective forces affected genes evolution. First, advantageous alleles that have<br />

swept on one si<strong>de</strong> of the barrier (unidirectional positive selection) are likely to cross this<br />

barrier faster than non-advantageous ones to inva<strong>de</strong> the entire population, also dragging<br />

along a series of linked neutral mutations by genetic hitchhiking (selective sweep, Faure et<br />

127


Chapitre 4<br />

al 2008). In a similar way, balancing selection favours introgression for alleles associated<br />

with balanced polymorphisms in recently diverged species for which genetic isolation is<br />

incomplete (Castric et al 2008). The expected consequence of positive selection acting on a<br />

gene, either un<strong>de</strong>r directional or balanced forces is therefore a <strong>de</strong>crease of genetic<br />

differentiation in structured populations when compared to neutral genes simply because<br />

effective migration rates of advantageous alleles are favoured (Muirhead, 2001, Glémin et<br />

al 2005). While unidirectional positive selection (selective sweep) tends to reduce genetic<br />

diversities, balancing selection promotes the allele diversification (Glémin et al 2005,<br />

Castric & Vekemans 2007, Castric et al 2008). As previously stated by Boon et al (2009),<br />

adaptive introgression or balanced polymorphisms can be <strong>de</strong>tected from introgressed<br />

genomes when looking at a semi-permeable barrier if the time of isolation has been longer<br />

enough. Second, bidirectional (disruptive) selection (i.e. positive selection affects alleles<br />

differently on each part of the barrier) is known to reinforce genetic isolation in allopatry<br />

(see Templeton 1981) and often lead to the rapid fixation of gametic incompatibilities.<br />

When the barrier to gene flow becomes more permeable, allowing secondary contacts<br />

between previously isolated <strong>de</strong>mes, the range of the hybrid zone will mainly <strong>de</strong>pend on the<br />

strength of selection against putative hybrids (i.e. the number of gametic incompatibilities<br />

between the two genomes). Such an evolutionary process will therefore have to slow down<br />

allele migrations around the barrier leading to either mosaic hybrid zones if dispersal is<br />

high or an abrupt tension zone if dispersal is restricted (See Bierne et al 2003).<br />

Consequently, because of differing evolution rates and selective impacts on genes,<br />

the use of both mitochondrial (four times smaller effective population size (Birky et al<br />

1983, 1989), and a higher evolutionary rate when compared to nuclear genes leading to the<br />

rapid fixation of mutations in the divergence (Neigel & Avise 1993, Irwin 2002)) and<br />

nuclear genes (allowing the access to hybridization process) is of great interest to better<br />

un<strong>de</strong>rstand gene histories and to assess more accurately the strength of such barriers on<br />

population structure (e.g. Godinho et al 2008, Burton & Lee 1994, Faure et al 2009).<br />

Deep-sea hydrothermal vents represent a patchy distributed habitat following a onedimensional<br />

stepping-stone mo<strong>de</strong>l associated with over 60 000 kilometres of globeencircling<br />

ridge crests (Schlager 2008). Because of their closely <strong>de</strong>pen<strong>de</strong>nce to ‘hot’<br />

sulphidic fluids from vents chimneys, <strong>de</strong>ep-sea hydrothermal vent species displayed an<br />

island-like repartition along ridges and raised questions about their inter-site dispersal and<br />

the putative occurrence of physical barriers to dispersal (Vrijenhoek et al 1997,<br />

128


Chapitre 4<br />

Chevaldonné et al 1997). According to the Kimura & Weiss (1964)’s one-dimensional<br />

stepping-stone mo<strong>de</strong>l, hydrothermal vent species were thus hypothesised to disperse<br />

primarily in their neighbourhood (Lutz 1988) and expected to follow an isolation-bydistance<br />

mo<strong>de</strong>l of dispersal (Vrijenhoek et al 1997). However, in most cases, allozyme<br />

studies failed to <strong>de</strong>tect isolation-by-distance across populations of hydrothermal vent<br />

species (e.g. Black et al 1994; Craddock et al 1995), suggesting that extensive gene flow<br />

better fit an n-island mo<strong>de</strong>l (Wright 1931).<br />

Amongst vent species, the <strong>de</strong>ep-sea hydrothermal vent polychaete Alvinella<br />

pompejana, which lives on the top of vent chimneys of the East Pacific Rise (EPR) also<br />

exhibited a lack of clear allozyme structure from latitu<strong>de</strong>s 9°50N to 21°N <strong>de</strong>spite a chaotic<br />

microgeographical structure at the field scale (i.e. tens of kilometres) (Jollivet et al 1995).<br />

These results indicated that larval dispersal is allowing high levels of gene exchanges<br />

providing that no physical barrier to gene flow disrupts the ridge system (Jollivet 1996).<br />

This fits well previous results of Pradillon et al (2001) who experimentally reported that A.<br />

pompejana embryos are able to stop their <strong>de</strong>velopment in cold water masses and thus to<br />

<strong>de</strong>lay their metamorphosis while dispersing out of vents. However, a propagule-flux mo<strong>de</strong>l<br />

based on Alvinella’s biology indicated that effective migration rate cannot be maintained<br />

across vent fields if larvae travel more than 15-30 days in the water column driven by the<br />

bottom currents prevailing on to the rift axis (Chevaldonné et al 1997). Consequently,<br />

according to the present ridge hydrodynamism, A. pompejana was still believed to disperse<br />

in the stepwise manner. To explain this absence of isolation-by-distance at the scale of the<br />

North EPR, three hypotheses have been proposed: (1) the lack of power of allozymes to<br />

<strong>de</strong>tect genetic differentiation between populations, (2) the maintenance of equal allozyme<br />

frequencies by balancing selection due to the great homogeneity of the vent conditions<br />

along the whole ridge system. or (3) the maintenance of gene exchanges within a large<br />

metapopulation by the displacement of habitat patches over geological times (Jollivet<br />

1996). The second hypothesis was exemplified by Piccino et al (2004) who reported a case<br />

of balancing selection on the phosphoglucomutase (PGM) with allele 90 being more<br />

thermostable and more frequent in newly foun<strong>de</strong>d populations of the worm but balanced at<br />

the metapopulation scale. The latest hypothesis was also tested by mo<strong>de</strong>lling the<br />

hydrothermal vent activity movements along the ridge and its subsequent effect<br />

(extinction/recolonisations) on the genetic structure of the worm over thousands of<br />

generations (Jollivet et al 1999). Simulations showed that bursts of colonisation events<br />

following site movements and population reconnections were sufficient enough to prevent<br />

129


Chapitre 4<br />

population isolation and thus genetic differentiation between distant sites, providing that<br />

worms are not able to disperse very far to re-colonise nearby sites. The mo<strong>de</strong>l predictions<br />

were however based on to the fact that habitat displacements can cross physical barriers<br />

such as transform faults.<br />

Recently, a genetic divergence of nearly 1% was observed on the mitochondrial<br />

COI gene between the Northern and Southern EPR populations of A. pompejana (Hurtado<br />

et al 2004, Plouviez et al 2009), suggesting the occurrence of an impermeable equatorial<br />

barrier to gene flow. Furthermore, Plouviez et al (2009) <strong>de</strong>monstrated that this barrier<br />

affected most of the EPR vent fauna 1 to 2 Mya, giving cre<strong>de</strong>nce to the role of transform<br />

faults in provoking vicariant events between vent communities and allopatric speciation. In<br />

the present study, an exten<strong>de</strong>d analysis of allozyme and mtCOI haplotype distributions<br />

over the geographic range of the species, including the Southern EPR, was performed to<br />

better address the ability of allozymes to <strong>de</strong>tect genetic differentiation and to discriminate<br />

between the three main hypotheses that have been proposed to explain the lack of<br />

differentiation between alvinellid populations in the North. Three additional nuclear genes,<br />

including the PGM gene, were used to evaluate the strength of this equatorial barrier to<br />

gene flow and to test for potential secondary contacts. Analyses also attempted to quantify<br />

migration and/or introgression rates between the previously isolated mt lineages in or<strong>de</strong>r to<br />

disentangle the effect of both selective and <strong>de</strong>mographic processes in the raise of the<br />

genetic break.<br />

130


Chapitre 4<br />

Materials and Methods<br />

Collection<br />

Alvinella pompejana polychaetes were collected with the manned submersible Nautile<br />

during three oceanographic cruises in 1999 (HOPE), and 2002 (PHARE) on sites of the<br />

northern EPR and in 2004 (BIOSPEEDO) on sites along the southern EPR from the<br />

latitu<strong>de</strong>s 7°25S to 21°33S. Once aboard the ship (the N/O L’Atalante), samples were<br />

measured at the 4h setigerous segment, sexed and cut in two parts: the cephalic part was<br />

frozen immediately at -80°C to perform allozymes and the body was preserved in 95%<br />

ethanol for DNA analyses.<br />

Allozyme genotyping<br />

Allozymes enco<strong>de</strong>d by nine enzyme systems were examined from seven hydrothermal vent<br />

field samples (Table 4.1), following the protocols of Pasteur et al (1987). Proteins from<br />

each frozen gill tissue were extracted according to the Piccino et al (2004)’s method.<br />

Enzyme electrophoreses were conducted on 12% starch gel using three different buffer<br />

systems: (1) Tris-Citrate pH 8.0 (TC 8.0) for Phosphoglucomutase (Pgm, E.C. 5.4.2.2),<br />

Mannose phosphate isomerase (Mpi, E.C. 5.3.1.8), 6-phosphogluconate <strong>de</strong>hydrogenase<br />

(6Pgd, E.C. 1.1.1.44) and Glucose phosphate isomerase (Gpi, E.C. 5.3.1.9); (2) Tris-HCl<br />

pH 8.5/8.2 (THCl 8.5) for Triose phosphate isomerase (Tpi, E.C. 5.3.1.1), Leucine amino<br />

peptidase (Lap, 3.4.11.1) and Hexokinase-1 (Hk, 2.7.1.1); and (3) Tris-Citrate pH 6.7/6.3<br />

(TC 6.7) for acid phosphatase (Acp, E.C. 3.1.3.2) and Isocitrate <strong>de</strong>hydrogenase (Idh, E.C.<br />

1.1.1.42).<br />

131


Chapitre 4<br />

Table 4.1 Location, and sample sizes used in allozymes (n allo ), and mtCOI (n mtCOI ),<br />

and nuclear genes.<br />

n GlobX , n SAHH , n PGM : number of recaptured sequences for each nuclear gene using a<br />

maximum of two consensus sequences (alleles) per individual with, in bold: number of<br />

recaptured individuals, n RFLP : number of genotyped individuals in RFLP of PGM<br />

Vent field Geographic position Depth (m) n allo n mtCOI n GlobX n SAHH n PGM n RFLP<br />

13°N 12°43-50’N 2560-2700 40 67 19 17 20 74<br />

103°53-57’W<br />

10 11 12<br />

9°50N 9°31-51’N 2500-2585 0 10 7 20 8 0<br />

104°15-18’W<br />

4 15 6<br />

7°25S 7°25’S 2735-2752 17 14 14 13 15 12<br />

107°47-49W<br />

9 8 11<br />

14°S 13°59’S 2623-2632 15 20 16 16 15 16<br />

112°29’W<br />

9 14 11<br />

17°25’S 17°25’S 2578-2590 34 30 16 16 11 15<br />

113°12’W<br />

10 9 8<br />

17°34’S 17°35-36’S 2591-2597 19 36 12 15 12 17<br />

113°15’W<br />

18 10 10<br />

18°25S 18°26-37’S 2636-2680 52 29 17 17 14 14<br />

113°23-24’W<br />

11 14 10<br />

21°25’S 21°25’S 2804-2831 12 30 16 16 13 22<br />

114°16-17’W<br />

12 11 10<br />

21°33’S 21°33’S<br />

114°16-18’W<br />

2771-2846 23 25 14<br />

10<br />

20<br />

14<br />

17<br />

13<br />

11<br />

DNA sequencing<br />

Genomic DNA was extracted using a CTAB extraction procedure. Tissues were<br />

digested in 600 µl of a 2% CTAB buffer solution (1.4 M NaCl, 0.2% 2-mercaptoethanol,<br />

20 mM EDTA, 100 mM Tris-HCl pH 8, 0.1 mg.ml -1 proteinase K) during two hours at<br />

60°C. DNA was then purified by adding Chloroform-Isoamyl alcohol (24:1) and<br />

precipitated together with 1 ml of 100 ml of 100% isopropanol at -20°C for two hours.<br />

Finally, DNA pellets were washed with 70% ethanol and re-suspen<strong>de</strong>d in 50 µl sterile H 2 0.<br />

A. pompejana sequences of the mitochondrial Cytochrome c Oxidase subunit I (mtCOI)<br />

were obtained from Plouviez et al (2009) and the sequences of the outgroup A. caudata<br />

using the ‘universal’ LCOI and HCOI (Folmer et al 1994) primers following PCR<br />

conditions <strong>de</strong>scribed by these authors. Specific primers (Table 4.2) were <strong>de</strong>signed for<br />

DNA amplification of partial exon-intron fragment of three nuclear genes, coming from<br />

132


Chapitre 4<br />

the sequencing of the transcriptome of the southern mtCOI lineage of A. pompejana for the<br />

S-a<strong>de</strong>nosylhomocysteine hydrolase (SAHH) and globin (GlobX) genes (Gagnière et al<br />

submitted), and from the overall phosphoglucomutase gene of A. pompejana previously<br />

obtained using a <strong>de</strong>generated primers approach (D. Jollivet, unpublished data). This latter<br />

approach allowed us to target Pgm-exon4 where a non-synonymous change was found.<br />

Size of the DNA fragment and its exon-intron structure are given in Fig. 4.7.<br />

Table 4.2 Primers <strong>de</strong>signed for Alvinella pompejana amplification of mtCOI, GlobX,<br />

SAHH, and PGM genes and for the screening of in<strong>de</strong>l polymorphism of the PGM<br />

exon 4 gene.<br />

----: corresponds to the nucleoti<strong>de</strong> ‘-bp tag’ of primers for the Mark-Recapture sequence<br />

method.<br />

Gene Primer sequences (5’-3’) T a (°C)<br />

mtCOI<br />

F: TATTTGGTATTTGGGCAGGTC<br />

57<br />

R: GATGGGTCGAAGAATGATGTG<br />

GlobX<br />

F: ----AGTTGACTGAAGAACGCAGAGAAGCGGT 55<br />

R: ----CGCGAAGATCCTTGAAGTACTCTTGGTA<br />

SAHH<br />

F: ----CTTGTCAACCTCGGCTGTG<br />

57<br />

R: ----CTGCCTGCTCGTCTGTTAG<br />

PGM<br />

F: ----CCCAGGTGGTCCAAATGC<br />

57<br />

R: ----GTACAGATGTCAGCCTTCAGGTC<br />

In<strong>de</strong>l polymorphism F, F*: GCCATGTTTTTAATTGATTCTGTC<br />

58<br />

in PGM exon4 R: TCACTACAAAATGAGCTGATGTGC<br />

F, forward primer; F*, forward labelled primer IRD700; R, reverse primer; T a , annealing<br />

temperature<br />

133


Chapitre 4<br />

Fig. 4.7 Comparison of size and exon-intron structure of the four sequenced<br />

genes<br />

Polymerase chain reactions (PCR) and cloning of these gene fragments were performed<br />

using the Mark-Recapture (MR) method <strong>de</strong>veloped by Bierne et al (2007) with 4-<br />

nucleoti<strong>de</strong>s tail ad<strong>de</strong>d at the end of each 5’-primers. Each individual was amplified by PCR<br />

separately with a different combination of 5’ tailed primers. PCR were conducted in a 25-<br />

µl solution that inclu<strong>de</strong>d 1X buffer (supplied by manufacturer), 2 mM MgCl 2 , 0.05 mM<br />

(0.25 mM for PGM) of each dNTPs, 0.48 µM (0.4 µM for PGM) of each primers (Table<br />

4.1), 0.5 U of Taq polymerase (Thermoprime plus), 5 µl of template DNA and sterile H 2 O.<br />

Thermal cycling parameters used an initial <strong>de</strong>naturation at 94°C for 3 min, followed by 40<br />

cycles at 94°C for 30s, T a for 20s and 72°C for 2min30s, before a final 10 min extension at<br />

72°C. Because of the MR protocol, eight PCR-products were mixed together per cloning<br />

with two clonings per vent locality (16 individuals per sampled site). PCR-products were<br />

isostochiometrically mixed together, purified using QIAQuick TM columns and ligated into<br />

a pGEM-T vector (pGEM-T cloning kit, Promega, Madison, WI, USA). For each cloning,<br />

32 positive clones were sequenced both strands at the Genoscope (Evry) with universal<br />

primers of the plasmid (SP6 and T7), leading to a total number of around 3500 nuclear<br />

sequences. Sequences from all genes were proofread using CodonCo<strong>de</strong> Aligner 2.0.6<br />

134


Chapitre 4<br />

(http://www.codonco<strong>de</strong>.com/aligner/). Sequence alignments were initially performed with<br />

ClustalW (Thompson et al 1994) in BioEdit version 6.0.6 (Hall, 1999) and subsequently<br />

improved manually. For nuclear genes, in vitro recombinants between individuals from the<br />

same cloning set were easily <strong>de</strong>tected thanks to their abnormal combination of 5’-tails and<br />

removed from the dataset. Multiple recaptures also allowed us to discard intra-individual in<br />

vitro recombinants (PCR recombination of heterozygous alleles) and artefactual/somatic<br />

mutations.<br />

Detecting a putative barrier to gene flow<br />

Linkage disequilibrium between allozymes was examined for each locality and over<br />

the all dataset using Genetix version 4.05.2 to <strong>de</strong>tect redundant information. For each<br />

locality, allele frequencies were then calculated for each locus and <strong>de</strong>parture from Hardy-<br />

Weinberg equilibrium was estimated using the Weir & Cockerham (1984) f estimator and<br />

tested by 1000 permutations. To <strong>de</strong>tect putative barrier to gene flow, genetic differentiation<br />

was estimated along the EPR from a multiloci Fst as calculated by the Weir &<br />

Cockerham’s θ (1984), adding populations one by one from the most southern locality<br />

21°33’S to 13°N. At each step, the F in<strong>de</strong>x was tested by 1000 permutations of the dataset.<br />

Isolation-by-distance was tested along the whole EPR using a Mantel Z test (Mantel 1967)<br />

with 5000 permutations. Because previous studies (Hurtado et al 2004; Plouviez et al<br />

2009) <strong>de</strong>tected a barrier to gene flow across the Equator for the mtCOI marker, this test<br />

was also performed without using the 13°N population. To better un<strong>de</strong>rstand the<br />

population structure, the number of genetically-differentiated groups (k) was evaluated<br />

using individual assignments to these a priori groups based on their multiloci genotype<br />

from an admixture mo<strong>de</strong>l with a 150 000 burnin length period following by 50 000 MCMC<br />

repetitions. The most appropriate number was obtained by testing k values from 1 to 15<br />

thanks to Structure version 2.2. (Pritchard et al 2000). Convergence of the Markov chain<br />

simulations was checked toward the true stationary distribution by monitoring ten<br />

iterations.<br />

For all statistical analyses based on recaptured sequences, one consensus allele was<br />

randomly selected per individual to avoid the putative bias due to the cloning method.<br />

Allele diversity Hd, nucleoti<strong>de</strong> diversity π N , and theta of Watterson θ W were estimated<br />

from insi<strong>de</strong> the northern and southern EPR separately with DNAsp 4.10.3 (Rozas et al<br />

135


Chapitre 4<br />

2003). For each gene, allele networks were constructed using Network 4.5.1.0<br />

(www.fluxus-engineering.com: Ban<strong>de</strong>lt et al 1999) to search for divergence across the<br />

barrier to gene flow (all recaptured alleles were used). Genetic differentiation in<strong>de</strong>xes φ st<br />

were computed via Arlequin 3.1 (using nucleoti<strong>de</strong> sites: Excoffier et al 2005) by grouping<br />

populations from each part of the equator and <strong>de</strong>partures from zero (no differentiation)<br />

were tested by a 1000-permutation test.<br />

Because most coalescent-based softwares impose strong assumptions about the<br />

absence of recombination, recombinants were <strong>de</strong>tected by the Hudson & Kaplan (1985)’s<br />

four-gamete test, using DNAsp 4.10.3, and discar<strong>de</strong>d from the dataset for the following<br />

analyses: Migrate_n, IM/IMa and Sweep_bott.<br />

Detection of gene flow between South and North EPR across the equator<br />

Migration routes across the equator were examined via Migrate 3.0.3 (Beerli &<br />

Felsenstein 2001) by looking at the two geographically differentiated groups of sites<br />

(northern and southern parts of the geographical barrier). Unfortunately, because of the<br />

random sampling of alleles within each gene associated with the MR method, very few<br />

individuals (3.5 on average per population) displayed alleles in all genes. Migration rates<br />

and thetas were thus estimated by MCMC method for each gene, separately. Watterson’s<br />

theta (θ w ) were used as a starting parameter in Migrate 3.0.3 to estimate the gene flow<br />

parameters (θ*M) with MCMC runs of 10 short (500 steps, 10 000 sampled genealogies),<br />

three long chains (5 000 steps, 100 000 sampled genealogies) and swapping procedure<br />

among 4 heated chains.<br />

Detection of gene flow between the two mitochondrial cla<strong>de</strong>s of A. pompejana using an<br />

Isolation with Migration mo<strong>de</strong>l<br />

Population sizes, migration rate and time since population splitting were estimated<br />

using the Isolation with Migration mo<strong>de</strong>l (IMa software, Hey & Nielsen 2007). This mo<strong>de</strong>l<br />

allowed us to calculate posterior <strong>de</strong>nsity probabilities of these parameters assuming that<br />

migration could still occur after the splitting event of an ancestral population into two<br />

<strong>de</strong>scendant ones. To check convergence of the Markov Chain Monte Carlo (MCMC)<br />

toward the true stationary distribution, multiple runs were performed using different<br />

starting points and autocorrelation between parameter values was assessed over the course<br />

136


Chapitre 4<br />

of the runs. A swapping procedure was also used to enhance the mixing of chains using a<br />

geometric heating scheme with 6 Metropolis chains. Inheritance scalar was assigned to 1<br />

for nuclear genes (autosomal) and 0.25 for mtCOI to adjust for their expected population<br />

sizes. The Hasegawa-Kishino-Yano mo<strong>de</strong>l (Hasegawa et al 1985) was chosen to allow for<br />

multiple substitutions at the same site.<br />

To better estimate the date since the splitting event, IMa parameters of divergence<br />

between populations (t IMa ) have been converted in years. Because mutation rate has not<br />

been previously estimated for alvinellid polychaetes, with the exception of the mtCOI gene<br />

(Chevaldonné et al 2002) which is known to display a higher mutation rate when compared<br />

with nuclear genes, a geometric mean has been <strong>de</strong>rived from COI and PGM and SAHH for<br />

which a mutation rate was previously obtained from the hydrothermal vent gastropod<br />

Lepetodrilus fucensis (Johnson et al 2006) and the hydrothermal vent mussels<br />

Bathymodiolus thermophilus (Faure et al 2009) respectively. The GlobX gene was not<br />

taken into account since there is no information on its evolutionary rate so far. Before<br />

calculating the geometric mean, an average mutation rate was assessed per locus per<br />

generation accounting for different lengths of the 3 genes (COI 507 bp, PGM 940 bp,<br />

SAHH 531 bp) assuming one generation per year. The geometrical mean U = 0.64 × 10 -6 ,<br />

close to the value previously found using 8 genes (mtCOI and 7 nuclear genes, U = 0.63 ×<br />

10 -6 ) on B. thermophilus (Faure et al 2009), was then used to recalibrate the divergence<br />

time from IMa to time in years (T = t IMa / U).<br />

Detecting putative hybrids between the North and the South EPR using the PGM gene<br />

Because a substantial divergence has been observed between the northern and<br />

southern EPR cla<strong>de</strong>s on the nuclear PGM gene, fixed mutations were used to <strong>de</strong>tect and<br />

locate ‘hybrids’ between these two geographical regions (heterozygous individuals with<br />

one allele from Northern cla<strong>de</strong> and one allele from the Southern cla<strong>de</strong> or individuals<br />

sampled in a geographic region with alleles typical of the cla<strong>de</strong> of the other region). A<br />

restriction fragment length polymorphism (RFLP) analysis together with an in<strong>de</strong>l<br />

genotyping was performed on the polychaete collection to access a composite PGM<br />

genotype of each worm. Among the 4 cla<strong>de</strong>s observed, cla<strong>de</strong> 3 displayed two restriction<br />

sites for the Rsa1 enzyme instead of one for the three remaining cla<strong>de</strong>s. Cla<strong>de</strong> 4 displayed<br />

two restrictions sites for Ssp1 instead of three sites for the three remaining cla<strong>de</strong>s. Cla<strong>de</strong>s 1<br />

and 2 were different from each other by several fixed substitutions for which no restriction<br />

137


Chapitre 4<br />

site was observed and the presence of a nine bp in<strong>de</strong>l. RsaI and Ssp1 RFLP were thus<br />

performed on PGM PCR-products following protocols of the manufacturers, and the<br />

presence/absence of the in<strong>de</strong>l was genotyped using new nested primers <strong>de</strong>fined from each<br />

part of this in<strong>de</strong>l. Individuals were amplified in 15-µl solution that inclu<strong>de</strong>d 1x buffer<br />

(supplied by manufacturer), 1.25 mM MgCl 2 , 0.2 mM each dNTPs, 0.1 µM of fluorescent<br />

labelled (IRD700) forward primer, 0.27 µM of unlabelled forward primer, 0.33 µM of<br />

unlabelled reverse primer (Table 4.1), 0.5 U of Taq polymerase (Thermoprime plus), 4 µl<br />

of template DNA and sterile H 2 O. Thermal cycling parameters used an initial <strong>de</strong>naturation<br />

at 94°C for 2 min, followed by 30 cycles at 94°C for 30s, annealing temperature (T a , Table<br />

4.1) for 30s and 72°C for 25s, before a final 10 min extension at 72°C. PCR products were<br />

then electrophoresed on a <strong>de</strong>naturing acrylami<strong>de</strong> 41 cm gel in a Li-Cor NEN Global IR2<br />

DNA analyzer.<br />

1<br />

Detecting <strong>de</strong>mographic and/or selective events from coalescence trees<br />

Because topologies of coalescent can help to perceive <strong>de</strong>mographic and/or selective<br />

events, Neighbour Joining trees (distances obtained by the maximum composite<br />

likelihood) were performed using Mega 4.0.2 (Tamura et al 2007), separately on North and<br />

South mt cla<strong>de</strong>s and rooted with an A. caudata sequence for non-ambiguously well-aligned<br />

portions of each gene. Departure from the neutral hypothesis was assessed by estimating<br />

Tajima (D) and Fu & Li (F*) in<strong>de</strong>xes in each cla<strong>de</strong> to test for putative <strong>de</strong>mographic and/or<br />

selective events. A multi-locus Hudson-Kreitman-Aguadé test (HKA, 1987) was computed<br />

for each A. pompejana lineages with one sequence of A. caudata as outgroup using the<br />

software HKA (J. Hey’s web page:<br />

http://lifesci.rutgers.edu/∼heylab/HeylabSoftware.htm#HKA). This test compares<br />

divergence and polymorphism at several loci to <strong>de</strong>tect if some of these loci display<br />

<strong>de</strong>parture to neutral evolution. McDonald Kreitman tests (MK) were then performed<br />

between these two species to test for possible selection in the divergence 1 . Discrimination<br />

between bottleneck versus selective sweep was tested from a series of 4 genes (including<br />

mtCOI) both in Northern and Southern mt cla<strong>de</strong>s, after excluding introgressed alleles,<br />

using sweep_bott (Galtier et al 2000) by 100 000 first step iterations, 1 000 000 second<br />

iterations and 20 optimizations processes with a theta range from 1 to 30. This test is<br />

1 Le co<strong>de</strong> génétique mitochondrial <strong>de</strong> Drosophila a été utilisé <strong>pour</strong> le mtCOI : cf annexe 4<br />

138


Chapitre 4<br />

comparing 3 distinct coalescent mo<strong>de</strong>ls (including the neutral one) and estimates whether<br />

the strength and the time since bottleneck are sufficiently close one to each other across<br />

genes to cope with a real bottleneck.<br />

139


Chapitre 4<br />

Results<br />

Equatorial barrier to gene flow<br />

Linkage disequilibria have not been found significant for all loci pairs over all<br />

populations (p > 0.05 for the Garnier-Géré & Dillman (1992) test), indicating that locus<br />

provi<strong>de</strong>d in<strong>de</strong>pen<strong>de</strong>nt information. The number of alleles ranged from 3 (GPI, IDH, ACP,<br />

LAP, HK) to 5 (PGM, TPI, see: Supplementary material). However, most loci displayed<br />

only 1 to 3 non-rare alleles (frequency > 5% in the total population), which were shared<br />

between populations. In contrast, rare alleles were often private (Supplementary material).<br />

Alleles displayed a frequency inversion only for PGM and TPI loci between 13°N and<br />

7°25S (Fig. 4.8). The allele PGM 100 was more frequent at 13°N than in southern EPR<br />

populations where allele 90 predominates. A similar situation was <strong>de</strong>picted at the TPI<br />

locus for which allele 110 predominates at 13°N whereas it was allele 100 in the southern<br />

populations. Significant heterozygote <strong>de</strong>ficiencies (Fis > 0, see Supplementary material)<br />

were not linked to a particular locus and not systematically found at a locality. However, a<br />

more accurate examination of the Fis distribution between loci that yield two-to three<br />

equifrequent alleles indicated that the highest values were mostly found at 14°S and<br />

17°25S.<br />

140


Chapitre 4<br />

Fig. 4.8 Allozyme frequency clines along the East Pacific Rise for the PGM, MPI,<br />

TPI, HK and 6PGD loci.<br />

The multilocus estimate of Weir & Cockerham’s θ (1984) overall populations was<br />

significantly different from 0 (θ = 0.134, p < 0.05, 1000 permutations). This in<strong>de</strong>x of<br />

fixation estimated overall loci by adding populations one by one from 21°33’S to 13°N<br />

became significant only when adding the 13°N population (θ = 0.091, p < 0.05). Even if a<br />

trend of isolation-by-distance exists (see: Fig 4.9), the correlation between the genetic<br />

distance θ(1-θ) and the geographic distance was not significant (Mantel Z test: p = 0.21;<br />

1000 permutations) at the scale of the whole EPR and not even more by removing 13°N<br />

from the dataset (p = 0.20).<br />

141


Chapitre 4<br />

p = 0.21<br />

R 2 = 0.74<br />

Fig. 4.9 Distribution of genetic differentiation according to distance.<br />

P-value corresponds in the Mantel Z test result and R 2 to the correlation coefficient in the<br />

overall EPR. Black squares: 17°25’S to 21°33’S. Black squarse + white circles: Southern<br />

EPR. All symbols: the overall EPR.<br />

142


Chapitre 4<br />

Partitioning genetic structure into groups using the Bayesian clustering method of<br />

the Structure 2.2 software (Pritchard et al 2000) that uses multiloci genotype assignments<br />

displayed 3 genetically different groups, found at different proportions within each vent<br />

field. Individuals from the 13°N vent field mainly accounted for the first group whereas<br />

7°25’S-14°S individuals were mostly found in the second group and 17°25’S to 21°S<br />

individuals in the third group. Because the group of 7°25’S-14°S may represent a transition<br />

zone between the Northern and Southern EPR, isolation-by-distance was also tested on a<br />

more restricted area (17°25’S-21°S) without success (Mantel test p = 0.64, see: Fig. 4.9).<br />

Sequence polymorphisms of 3 nuclear genes (SAHH, GlobX and PGM) and the<br />

mitochondrial gene COI were analysed from each part of the Equator. The summary<br />

statistics Hd, π N and θ W are presented in Table 4.3.<br />

Table 4.3 Gene diversity of Alvinella pompejana.<br />

n, number of sequences; Hd, Haplotype diversity, π n nucleoti<strong>de</strong> diversity; θ w , Watterson’s<br />

theta per site from the number of segregating sites, D, Tajima’s D in<strong>de</strong>x; F*, Fu et Li<br />

in<strong>de</strong>x; SD, standard <strong>de</strong>viation; for D, NS :P > 0.10<br />

Gene Locality n Hd (SD) π N (SD) θ W (SD) D F*<br />

mtCOI<br />

Cla<strong>de</strong> North 108 0.911 (0.014) 0.009 (0.000) 0.014 (0.004) -0.983 NS -2.164 NS<br />

GlobX<br />

SAHH<br />

PGM<br />

Cla<strong>de</strong> South 210 0.618 (0.040) 0.002 (0.000) 0.015 (0.004) -2.564 *** -4.330 **<br />

North 14 0.890 (0.060) 0.003 (0.000) 0.0034 (0.0017) -0.872 NS -1.504 NS<br />

South 73 0.860 (0.031) 0.0054 (0.001) 0.012 (0.004) -1.721 + -4.107 **<br />

North 26 0.471 (0.119) 0.002 (0.001) 0.004 (0.002) -1.989 * -2.7696 *<br />

South 80 0.604 (0.064) 0.0031 (0.001) 0.011 (0.003) -2.202 ** -4.484 **<br />

Cla<strong>de</strong>s North 18 0.921 (0.030) 0.0056 (0.001) 0.007 (0.003) -0.693 NS -1.437 NS<br />

Cla<strong>de</strong>s South 74 0.725 (0.055) 0.0042 (0.001) 0.009 (0.003) -1.836 * -3.950 **<br />

143


Chapitre 4<br />

As previously shown in Plouviez et al (2009), networks of mtCOI displayed two<br />

monophyletic reciprocal 1%-divergent cla<strong>de</strong>s in the Northern and Southern EPR,<br />

respectively (Fig. 4.10). Similarly, PGM gene showed two sets of two divergent cla<strong>de</strong>s that<br />

were geographically structured (average North/South divergence = 0.76%): cla<strong>de</strong>s 1 and 2<br />

were mainly found in the northern EPR whereas cla<strong>de</strong>s 3 and 4 were distributed in the<br />

southern EPR (Fig. 4.10). Populations from the Northern and Southern EPR thus showed<br />

φ st values significantly different from zero at this locus (φ st = 0.18, p = 0), indicating the<br />

occurrence of a significant genetic differentiation between these regions. On the contrary,<br />

the SAHH and GlobX genes provi<strong>de</strong>d a lack of geographically-structured coalescents<br />

without any real divergence between the North and South parts of the EPR. If populations<br />

appeared to be genetically homogeneous across the equator at the SAHH locus (φ st = 0, p ><br />

0.05), these populations were found genetically differentiated between these two regions at<br />

the GlobX locus with significant φ st value (= 0.04, p < 0.05).<br />

144


Chapitre 4<br />

COI<br />

PGM<br />

2<br />

Ac<br />

1<br />

3<br />

4<br />

SAHH<br />

GlobX<br />

Ac<br />

Fig. 4.10 Neighbour Joining networks of the mtCOI and nuclear genes. Blue and red<br />

labels represent individuals sampled in Northern and Southern EPR, respectively.<br />

145


Chapitre 4<br />

Migration across the equator<br />

When estimating migration across the equatorial barrier from the two groups of<br />

geographically-pooled populations using the software Migrate_n 3.0.3 (Beerly &<br />

Felsenstein 1999), migration was clearly asymmetric from very low (PGM) to high levels<br />

of gene flow (SAHH) from North to South for all nuclear genes whereas a low (one<br />

southern haplotype sampled in the northern region) but significant migration was observed<br />

for the mitochondrial gene in the opposite direction (Table 4.4). However, migration rates<br />

obtained for both the mtCOI and PGM were close to zero indicating that the barrier is<br />

nearly impermeable for these two genes.<br />

Table 4.4 Migration rates across the equatorial barrier using the software Migrate_n.<br />

N/S, North to South; S/N, South to North; θ N , Theta in the North; θ S , Theta in the South;<br />

MLE, Maximum-likelihood Estimates; P, percentile; -, estimated values of 4Nm failed for<br />

the consi<strong>de</strong>red percentile (too much lower values)<br />

Loci 4Nm N/S 4Nm S/N θ N θ S<br />

COI<br />

MLE 2.235 10 -13 0.572 0.016 0.024<br />

P 0.005 - 0.045 0.012 0.018<br />

P 0.995 0.781 2.209 0.024 0.031<br />

SAHH<br />

MLE 23.389 1.302 10 -13 0.005 0.016<br />

P 0.005 13.815 - 0.004 0.011<br />

P 0.995 36.030 0.644 0.008 0.025<br />

GlobX<br />

MLE 10.359 1.007 10 -7 0.004 0.009<br />

P 0.005 5.014 - 0.002 0.006<br />

P 0.995 16.830 - 0.006 0.014<br />

PGM<br />

MLE 1.525 4.546 10 -13 0.011 0.010<br />

P 0.005 0.326 - 0.003 0.006<br />

P 0.995 3.872 0.855 0.022 0.022<br />

146


Chapitre 4<br />

Testing an Isolation Mo<strong>de</strong>l with Migration between the two mitochondrial cla<strong>de</strong>s of A.<br />

pompejana<br />

The mo<strong>de</strong>l of Isolation with Migration <strong>de</strong>veloped by Hey and Nielsen (2007) was<br />

used for testing the occurrence of either past gene flow between the two divergent mt<br />

lineages of A. pompejana or <strong>de</strong>tecting recent hybridization and introgression processes<br />

following a secondary contact zone. These migration rates are clearly different from those<br />

estimated by Migrate_n across the barrier since this later quantifies both the introgression<br />

rate between the two mt cla<strong>de</strong>s and the exchanges of individuals from both types across the<br />

barrier. The estimated ‘Isolation with migration’ parameters are presented in Table 4.5.<br />

Divergence time between the two mt lineages displayed the highest likelihood values<br />

around 1-2 Mya (Fig. 4.11). General patterns of allele migration occurred from the<br />

Southern to the Northern mtCOI cla<strong>de</strong> whereas no migration has been <strong>de</strong>tected in the<br />

opposite direction (Fig. 4.11, Table 4.5).<br />

147


Chapitre 4<br />

148


Chapitre 4<br />

A<br />

B<br />

C<br />

Fig. 4.11 Marginal posterior probability distribution for divergence time between<br />

lineages (A), effective population sizes of the ancestral population and its two<br />

<strong>de</strong>rived ones (B) and directional gene-flow rates between the two mt lineages (C),<br />

estimated by the isolation with migration mo<strong>de</strong>l IMa.<br />

Parameters correspond in those from the Table 4.5.<br />

149


Chapitre 4<br />

This result is concordant with the monolocus asymmetric direction from South to North of<br />

the migration (presence of a non null peak), although the probability distributions of<br />

migration for genes PGM, SAHH and GlobX are not significantly shifted from zero on the<br />

basis of the 90%HDP. This migration orientation can be clearly seen from the allele<br />

admixtures observed in the northern cla<strong>de</strong> when looking at the allele networks, at least for<br />

the PGM and GlobX genes (Fig. 4.12). The mean time of migration events from the<br />

Southern cla<strong>de</strong> to the Northern cla<strong>de</strong>s ranged from 25 000 to 300 000 years ago (Table<br />

4.6), indicating that alleles exchanges occurred closely after the isolation process (1.6<br />

Mya). The mean time of migration events (mtime) at PGM was however ten times lower<br />

than the two other loci, indicating that allele introgression occurred more recently at this<br />

locus (estimated around 23 000 years).<br />

Fig. 4.12 IM single locus probability distributions of migration rates for each of the<br />

4 loci.<br />

m1, migration from the Northern to Southern mtCOI lineages; m2, migration from<br />

Southern to Northern mtCOI lineages.<br />

150


Chapitre 4<br />

Table 4.6 Estimates of migration rates (m) and mean time of migration events<br />

(mtime), obtained from IM analyses on the 4 loci.<br />

Labels (1) and (2) correspond to the Northern and Southern cla<strong>de</strong>s, respectively. For<br />

SAHH and GlobX genes, m values correspond to the range of the posterior distribution<br />

without consi<strong>de</strong>ring m = 0.<br />

Locus<br />

m1 m2 mtime1 (My) mtime2 (My)<br />

COI 0 0 - -<br />

SAHH 0 0-7 - 0.347<br />

GlobX 0 0-7 - 0.388<br />

PGM 0 1 - 0.023<br />

Hybrid <strong>de</strong>tection between the North and the South cla<strong>de</strong>s of PGM<br />

The PGM network (Fig. 4.10) allowed us to <strong>de</strong>tect 3 individuals sampled on the<br />

Southern EPR with at least one allele of the Northern cla<strong>de</strong> (two from 14S and one from<br />

7°25S) but the Southern mitochondrial type. Because of fixed substitutions and in<strong>de</strong>ls into<br />

the divergence of the 4 PGM cla<strong>de</strong>s, all specimens collected were genotyped by combining<br />

RFLP and intron length. By comparing RFLP, in<strong>de</strong>l genotypes and mitochondrial<br />

haplotypes of each individual, 4 ‘hybrids’ (33.3%) between the Northern and Southern mt<br />

cla<strong>de</strong>s were <strong>de</strong>tected at 7°25S, 2 (16.7%) at 14°S and only one (1.7%) at 13°N whereas no<br />

‘hybrid’ has been found from the other latitu<strong>de</strong>s.<br />

Moreover, comparison between RFPL and allozymes allowed to <strong>de</strong>tect a clear<br />

correspon<strong>de</strong>nce between cla<strong>de</strong>s 3 and 4 <strong>de</strong>tected in PGM gene with allele 90 and 100<br />

found in allozymes in southern EPR, respectively (complete linkage disequilibrium:<br />

Garnier-Géré & Dillmann (1992) R’s correlation coefficient = 0.852, p = 0.0001).<br />

However, the 9bp in<strong>de</strong>l, together with linked substitutions that discriminate cla<strong>de</strong> 1 from<br />

cla<strong>de</strong> 2 in the Northern cla<strong>de</strong>, did not correspond to allozyme 100 and 90 of the northern<br />

populations. Three non-synonymous mutations were <strong>de</strong>tected within PGM exon4. Two of<br />

them were singleton and correspon<strong>de</strong>d to an Isoleucine / Valine amino acid replacement,<br />

having no or little influence on the protein charge. On the contrary, the third nonsynomynous<br />

mutation was a Gln/Glu change with a fixed Gln in cla<strong>de</strong> 3 (allozyme 90 in<br />

the South) and a fixed Glu for the three other cla<strong>de</strong>s. These two amino acids have not the<br />

same electrophoretic charge and are likely responsible for a net charge change of<br />

allozymes easily <strong>de</strong>tectable by electrophoretic migration on a starch gel. Because this non-<br />

151


Chapitre 4<br />

synonymous substitution was not <strong>de</strong>tected in PGM exon4 alleles of the Northern EPR<br />

populations, the amino acid substitution that is responsible for the 90/100 allelism in<br />

Northern EPR is elsewhere and represents an homoplasic site when compared with the<br />

alleles of exactly the same charge in the South.<br />

Detecting <strong>de</strong>mographic changes and/or selective effects<br />

Tajima (D) and Fu & Li (F*) in<strong>de</strong>xes were significantly negative for all loci in the<br />

Southern mt cla<strong>de</strong> whereas only the gene SAHH displayed significant values in the<br />

Northern cla<strong>de</strong>. All genes thus evi<strong>de</strong>nced a <strong>de</strong>parture from neutral equilibrium between<br />

drift and mutation in the Southern mt cla<strong>de</strong> but no necessarily in the Northern one.<br />

PGM-specific Neighbour Joining tree displayed a topology similar to gene un<strong>de</strong>r<br />

balancing selection (maintenance of ‘old’ lineages) but trees from the two other nuclear<br />

genes SAHH and GlobX had a more likely star-like shape typifying a selective sweep (Fig.<br />

4.13). The HKA multiloci test was significant using all genes (probability from χ 2<br />

distribution = 0.000), indicating that at least one gene evolved un<strong>de</strong>r positive selection.<br />

When discarding mtCOI, HKA test became not significant (pχ 2 = 0.134), suggesting that it<br />

could be the reflection of strong purifying selection at this locus. However, MacDonald<br />

Kreitman tests failed to <strong>de</strong>tect any selection whatever the gene consi<strong>de</strong>red (p > 0.05) by<br />

comparing A. pompejana polymorphism to the divergence to A. caudata outgroup. The<br />

Sweep_bott analyses performed within each mt cla<strong>de</strong>s using the 4 loci are summarized in<br />

Table 4.7.<br />

152


Chapitre 4<br />

Fig. 4.13 Linearized Neighbour Joining topologies of A. pompejana in easily aligned<br />

parts of the four studied genes in Northern and Southern EPR rooted by sequences<br />

from A. caudata (dotted lines).<br />

153


Chapitre 4<br />

Table 4.7 MLEs and likelihood ratio test (LRT) computed using the sweep_bott<br />

software to discriminate between the selective sweep and bottleneck hypotheses as<br />

shaping tree topologies of the Northern and Southern mt cla<strong>de</strong>s of A. pompejana.<br />

M1, neutral mo<strong>de</strong>l; M2, bottleneck mo<strong>de</strong>l; M3, selective sweep mo<strong>de</strong>l; NS, nonsignificant.<br />

Significances of the LRT were tested according to a χ2 table with a ddl equal<br />

to the difference between the number of parameters estimated for each mo<strong>de</strong>l (ddl = 2, 8<br />

and 6 for M2-M1, M3-M1 and M3-M2 multilocus LRT, respectively and ddl = 1 for M3-<br />

M1 monolocus LRT).<br />

MLE COI GlobX SAHH PGM<br />

North<br />

M1 -99.718 -37.728 -11.171 -14.341 -36.475<br />

M2 -92.660<br />

M3 -90.460 -35.171 -9.712 -11.170 -34.407<br />

LRT M2-M1: 14.116<br />

p < 0.005<br />

M3-M1: 18.516<br />

p < 0.025<br />

M3-M2: 4.400 (NS)<br />

5.114 (NS) 2.918 (NS) 6.342<br />

p < 0.025,<br />

4.136 (NS)<br />

South<br />

M1 -159.367 -4.149 -57.112 -38.064 -60.042<br />

M2 -153.211<br />

M3 -139.745 -0.290 -52.196 -30.557 -56.704<br />

LRT M2-M1: 12.312<br />

p < 0.005<br />

M3-M1: 39.243<br />

p < 0.005<br />

M3-M2: 23.282<br />

p < 0.005<br />

7.718<br />

p < 0.025<br />

9.830<br />

p < 0.01<br />

15.014<br />

p < 0.005<br />

6.677<br />

p < 0.005<br />

In the Northern cla<strong>de</strong>, likelihood ratio tests indicated that the bottleneck (M2 mo<strong>de</strong>l) and<br />

selective sweep (M3 mo<strong>de</strong>l) mo<strong>de</strong>ls were both significantly better to explain the coalescent<br />

topology than the neutral M1 mo<strong>de</strong>l. It was however not possible to discriminate between<br />

the M2 and M3 hypotheses. Likelihood-ratio tests (LRT) performed on each gene,<br />

separately were only significant for the SAHH gene, indicating that M3 mo<strong>de</strong>l would be<br />

more probable than M2 in the Northern cla<strong>de</strong>. In the Southern cla<strong>de</strong>, likelihood ratio tests<br />

also showed that M1 is not able to explain the observed tree topologies. Based on LRT,<br />

M3 was significantly better than M2, indicating that a selective sweep occurred at, at least<br />

one gene in the Southern cla<strong>de</strong>. However, all genes taken separately displayed a significant<br />

ratio test when compared with the neutral M1 mo<strong>de</strong>l, suggesting that a bottleneck might<br />

have occurred even if one gene (SAHH: greatest level of LRT significance across genes)<br />

may have swept across the two lineages. Consequently, analyses of <strong>de</strong>parture from neutral<br />

154


Chapitre 4<br />

evolution reinforce the hypothesis of population expansion in the southern EPR while also<br />

promoting the scenario of selective sweep that exten<strong>de</strong>d across the mt lineages of A.<br />

pompejana to explain the lack of geographic differentiation at the SAHH locus.<br />

Discussion<br />

Emergence and strength of an equatorial barrier to gene flow<br />

A genetic break for the mtCOI gene between the northern and southern EPR with a<br />

more or less exten<strong>de</strong>d transition zone for a series of hydrothermal vent species have been<br />

revealed in previous studies (Hurtado et al 2004, Plouviez et al 2009). If some species,<br />

such as the gastropod Lepetodrilus ovalis or the bivalve Bathymodiolus thermophilus,<br />

displayed a clinal distribution of their northern and southern mtCOI haplotypes (Plouviez<br />

et al 2009), Alvinella pompejana exhibited a clear equatorial break for this gene. In the<br />

present study, if three nearly-monomorphic enzyme loci did not show any genetic<br />

differentiation between the Northern and Southern EPR, the occurrence of an equatorial<br />

barrier was confirmed by the <strong>de</strong>tection of a clear genetic break at the six remaining enzyme<br />

loci. In addition, SAHH gene did not show significant differentiation across this barrier,<br />

but the GlobX and PGM genes displayed a slight allele-frequencies shift to nearly<br />

complete allele isolation. Variance of genetic differentiation among loci has been wi<strong>de</strong>ly<br />

reported during the allopatric isolation process (Lewontin & Krakauer 1973) and<br />

commonly <strong>de</strong>picted across a barrier to gene flow (Ting et al 2000, Broughton & Harrison<br />

2003, Pogson et al 1995, Lemaire et al 2005). This is mainly due to (1) the sensitivity of<br />

the molecular method used to <strong>de</strong>tect variants, which <strong>de</strong>pends on the time scale at which the<br />

isolation is observed (Sunnucks 2000, accumulation of non-synonymous mutations to<br />

change the net charge of a protein is a long process) and (2) the stochasticity of the<br />

coalescence process (Hudson & Turelli 2003).<br />

The strength of the barrier seems to be great as this barrier affected nearly all loci<br />

<strong>de</strong>spite is recent history. Divergence between northern and southern cla<strong>de</strong>s is in<strong>de</strong>ed<br />

estimated between 1-2 Mya using the IMa multilocus estimation with very low migration<br />

rates. This estimation is congruent with the shared vicariant event estimated to have<br />

occurred 1.3 Mya from a series of vent species using ABC computation on the<br />

mitochondrial gene COI (Plouviez et al 2009). The isolation process was believed to be<br />

155


Chapitre 4<br />

forced by the formation of multiple transform faults 1-2 Mya in this region (Kureth & Rea<br />

1981, Naar & Hey 1989, Francheteau et al 1990) and probably reinforced by a strong<br />

transverse <strong>de</strong>ep-sea current at the Equator (Reid 1997).<br />

Detection of a slight but significant differentiation for six enzyme loci across the<br />

equatorial barrier indicated that 1.3 My was not sufficient to affect the net charge of<br />

allozymes to create divergence but reduces migration enough to modify allozyme<br />

frequencies from each part of the barrier just by drift. Comparing patterns of genetic<br />

differentiation between allozymes and nuclear loci therefore suggests that even an abrupt<br />

<strong>de</strong>crease of gene flow due to a physical barrier to migration has a very limited impact in<br />

changing allozyme frequencies. However, this contrasts with the lack of genetic<br />

differentiation previously observed along the thousands kilometres stretch of both si<strong>de</strong>s of<br />

the EPR (Jollivet et al 1995, this present study) and confirms that A. pompejana is able to<br />

disperse farer than expected from the observed set of inter-field distances and the bottom<br />

current speed (Chevaldonné et al 1997). This situation may be therefore explained by<br />

either a <strong>de</strong>velopmental arrest of embryos in the cold abyssal water masses and subsequent<br />

<strong>de</strong>lay to metamorphosis (Pradillon et al 2001) or may also be imputed the displacement of<br />

sites due to hydrothermal activity movements allowing bursts of colonisation by<br />

reconnecting previously isolated populations (Jollivet et al 1999).<br />

Because the strength of a barrier to gene flow can vary through time, questions<br />

related to potential secondary contact during an ongoing allopatric isolation process may<br />

be explored in two ways: (1) is the barrier transiently permeable to gene flow or not<br />

anymore effective? and (2) are mtCOI lineages still able to crossbread and hybridize?<br />

Permeability of the barrier and a transient secondary contact zone<br />

A secondary contact zone between Northern and Southern EPR was already<br />

hypothesized to explain the clinal distribution of mtCOI for B. thermophilus and L. ovalis<br />

(Plouviez et al 2009). Such a secondary contact zone, if present, should be located near the<br />

barrier to gene flow with a size that would <strong>de</strong>pend on several parameters including the<br />

mo<strong>de</strong>l of dispersal of the species, the strength of the genetic barrier and the<br />

migration/selection balance on the gene un<strong>de</strong>r consi<strong>de</strong>ration (e.g. Piálek & Barton 1997).<br />

In<strong>de</strong>ed, Barton & Hewitt (1985) predicted that a genetic barrier is always sitting at a<br />

geographic point where gene flow falls to its minimum. It spatial extension mainly<br />

<strong>de</strong>pends on an equilibrium between the strength of the selection against hybrids and the<br />

156


Chapitre 4<br />

dispersal potential of propagules, leading to a narrow tension zone for weak dispersers<br />

(e.g. between 1775 and 1800m elevation in two Big Sagebruh subspecies, Graham et al<br />

1995) to large mosaic hybrid zones for long-term migrants (Bierne et al 2003).<br />

Estimation of migration rates across the equatorial barrier confirmed the<br />

existence of a narrow secondary contact zone, located between 7°25’S and 14°S. This<br />

supports the i<strong>de</strong>a that the barrier is porous. In<strong>de</strong>ed, if migration was quite inexistent for the<br />

gene mtCOI, all nuclear genes displayed an asymmetric migration flux across the equator<br />

from North to South. This was confirmed by the finding of some ‘hybrid’ individuals at the<br />

PGM locus at 7°25’S and 14°S, suggesting that the strength of the genetic barrier is still<br />

probably very high. Discrepancies of migration rates between loci could be due to<br />

selection against hybrids that can <strong>de</strong>lay the spread of allochtoneous alleles across the<br />

barrier and could thus affect genes differentially (Barton & Hewitt 1981, Barton &<br />

Bengtsson 1986). This suggests that allele introgression may occur between the mtCOI<br />

North/South lineages. However, the fact that gene flow is stopped for at least two loci<br />

(COI and PGM) raises the question on whether the barrier is becoming more and more<br />

impermeable with time: a situation that would fit the progressive offsetting of the ridge<br />

axis by a transform fault, or whether this barrier can be relaxed at some point to allow<br />

transient bursts of gene flow. Several biological explanations may support this hypothesis<br />

including the fact that larval exchanges across the barrier may be greatly facilitated by any<br />

modification of the <strong>de</strong>ep-sea current circulation prevailing at a given period of time<br />

(Hurtado et al 2004). Another possibility is that any displacement of hydrothermal vent<br />

discharge along the ridge axis may reduce the inter-field distance across the barrier enough<br />

to position populations at their closest proximity (Jollivet et al 1999).<br />

Fitting an Isolation with Migration mo<strong>de</strong>l (Hey & Nielsen 2004, 2007) between the<br />

mtCOI lineages may be helpful to estimate migration rates in the past and thus to predict<br />

whether these exchanges represent recent events or may have been much ol<strong>de</strong>r. Even if the<br />

estimated migration rates between the two cla<strong>de</strong>s were not significantly very different from<br />

zero, the shape of the posterior probability distribution of m2 (South to North) was shifted<br />

to a non-null estimate of migration for the three nuclear genes when compared to m1,<br />

confirming the South to North orientation of ‘migration’ (following the IM signification).<br />

This fits well with previous studies reporting that allele introgression is generally expected<br />

to occur in the opposite direction when compared to species dispersal and, concord with<br />

the hypothesis of selection against hybrids, which would tend to limit the spread of the<br />

new arriving alleles in the established lineage (Nagylaki 1975, Barton 1979, Barton &<br />

157


Chapitre 4<br />

Hewitt 1985). Mean times of migration between genes however indicated that migration<br />

between the two cla<strong>de</strong>s is probably not recent and may have been maximal around 20 to<br />

300 thousands years ago or more when compared to the 1-2 My phase of isolation. This<br />

indicated that secondary contact zone occurred during the formation of the barrier and<br />

allowed to some extent the hybridization of the two mtCOI lineages. It is therefore very<br />

difficult to discrimate the (even transient) secondary contact hypothesis from an ongoing<br />

speciation process in which gene flow is gradually reduced with time (see Nosil et al 2009,<br />

Faure et al 2009). In<strong>de</strong>ed, the isolation by “transform fault scenario” would tend to favour<br />

a gradual isolation process whereas secondary contacts could have been a direct<br />

consequence of a more stochastic process of speciation associated with the hydrothermal<br />

patch dynamics and local hydrodynamic changes.<br />

In addition to IM/IMa analyses, <strong>de</strong>tection of ‘hybrid’ individuals between the North<br />

and South cla<strong>de</strong>s using the PGM RFLP genotyping method provi<strong>de</strong>d evi<strong>de</strong>nce of a low<br />

level of geographically-restricted hybridization. These results are in accordance with the<br />

hypothesis of a tension zone between latitu<strong>de</strong>s 7°25’S and 14°S that may prevail for weak<br />

to mo<strong>de</strong>rate dispersers: a situation that also corroborates previous findings found with<br />

allozymes (i.e. significant heterozygote excesses and the Structure-based genetic<br />

peculiarities <strong>de</strong>tected in this narrow EPR region). On the opposite, absence of divergence<br />

between the Northern and Southern EPR cla<strong>de</strong>s for the SAHH and GlobX genes suggested<br />

either a wi<strong>de</strong>r introgression zone, possibly due to a lack of selection against hybrids at<br />

these loci, or the spread of an advantageous allele, which would have swept on one si<strong>de</strong> of<br />

the EPR and subsequently transferred in the other si<strong>de</strong>. Such differences in the span of<br />

introgressed alleles among loci are frequent in hybrid zones (Kruuk et al 1999). When<br />

consi<strong>de</strong>ring hybridizing species of cottonwoods, namely Populus fremontii and P.<br />

angustifolia along the Weber River, Martinsen et al (2001) already showed that about 20%<br />

of nuclear genes were differentially introgressed at distances from the contact zone varying<br />

from several km to over 100 km. A similar situation has been also <strong>de</strong>picted for long-term<br />

dispersing marine species (e.g. Bierne et al 2003 in Mytilus edulis and M. galloprovincialis<br />

mussels).<br />

158


Chapitre 4<br />

Testing the hypothesis of balancing selection as an homogenizing force across allozyme<br />

frequencies<br />

Allozymes failed to reveal isolation-by-distance neither along the Northern EPR (Jollivet<br />

et al 1995) nor across the Southern sites (present study). Using the software Structure<br />

populations were sub-divi<strong>de</strong>d in three genetically-differentiated groups mainly composed<br />

by the Northern EPR individuals, individuals from 7°25’S and 14°S and the most southern<br />

populations, respectively. However, although differentiated, these groups were not<br />

discriminated by the emergence of new allozymes and by the occurrence of three allele<br />

inversions. With the exception of TPI locus, the lack of an abrupt change in allele<br />

frequencies between 13°N and 7°25’S to 14°S together with non-significant pairwise Fst<br />

values may be a strong argument toward the assumption advocating that a strong<br />

directional selection driven by the geographic homogeneity of the vent habitat is acting on<br />

allozymes (Jollivet et al 1995). Genes un<strong>de</strong>r balancing selection or genes linked to them by<br />

hitch-hicking are known to preserve a greater population homogeneity than neutral ones in<br />

subdivi<strong>de</strong>d populations (Schierup et al 2000a, Schierup et al 2000b, Muirhead 2001).<br />

During the split of populations across a semi-permeable barrier, such genes should<br />

differentiate at a much lower rate than neutral genes. Looking at the sequence relationship<br />

between alleles encoding allozymes of the PGM, a locus previously supposed to be un<strong>de</strong>r<br />

balancing selection (Piccino et al 2004), was particularly interesting to <strong>de</strong>lineate the action<br />

of selection at this locus. Although the screening of the southern populations indicated an<br />

increase of allozyme 90 from North to South responsible for an overall significant genetic<br />

differentiation at this locus, the number of allozymes, their electrophoretic mobilities and<br />

subsequent frequencies remained nearly exactly the same from each part of the equatorial<br />

barrier (13°N/EPR: f 90 =0.4 and 7°25S: f 90 =0.55). Such a variation was not very different<br />

from variation found in A. pompejana between newly foun<strong>de</strong>d populations and more aged<br />

ones within the vent field 13°N/EPR (Jollivet et al 1995; Piccino et al 2004). However,<br />

looking at the divergence of alleles responsible for the enzyme polymorphism indicated<br />

that they are clearly distinct across the North/South barrier with a very small number of<br />

individuals possessing alleles from both mt lineages (“hybrids” mainly found at 7°25S).<br />

Moreover, allele assignments <strong>de</strong>rived from PGM exon4 (RFLP screening) to respectively<br />

allozymes 90 and 100, clearly showed that the Q/E substitution associated with the net<br />

charge difference of the southern allozymes do not correspond to the amino acid change<br />

159


Chapitre 4<br />

responsible for the same protein polymorphism in the Northern EPR. This situation was<br />

therefore puzzling and indicated that the electrophoretic mobility of allozymes 90 and 100<br />

was purely homoplasic between mt lineages and probably the result of an adaptive<br />

convergence rather than the maintenance of a balanced polymorphism across the barrier.<br />

Homoplasy in electrophoretic mobilities of allozymes has been wi<strong>de</strong>ly reported for<br />

incorrectly grouping species in phylogenies due to signal saturation (Keenan 1991, Zhang<br />

& Kumar 1997). However, very little information has been reported so far on homoplasy<br />

associated with allele electrophoretic mobility within a given species. Convergent<br />

evolution of alleles across distinct lineages was nevertheless reported at several loci (e.g.<br />

ABO gene: O’hUigin et al 1997 but see also Wood et al 2005). The shape of PGM<br />

coalescence trees and their associated neutrality tests however suggests that balancing<br />

selection may still act in the Northern populations (maintenance of two <strong>de</strong>eply divergent<br />

alleles without or little recombination) whereas it seems to behave more neutrally in the<br />

southern populations. The orientation of mutations of PGM alleles of A. pompejana by the<br />

outgroup A. caudata may help in the un<strong>de</strong>rstanding of the whole gene history. The network<br />

in<strong>de</strong>ed indicates that both alleles 90 and 100 from the southern populations <strong>de</strong>rived from<br />

only one of the two alleles found in the Northern populations, suggesting that (1) the<br />

northern alleles are probably more ancestral than the southern ones and (2) one ancestral<br />

allele at least was lost by drift during the colonisation of the Southern EPR. Such a<br />

scenario is highly predictable for populations subjected to high extinction rate and has been<br />

previously proposed to explain the absence of reciprocal monophilies at some genes during<br />

the process of allopatric speciation (Cunningham & Collins 1998).<br />

Possible impact of selective sweeps in the crossing of the physical barrier<br />

Balancing selection is not the only way to promote a lack of geographic structuring<br />

between populations. The rapid sweep of an advantageous allele in a given lineage is also<br />

expected to cross more easily a genetic barrier and spread within another lineage if the<br />

advantage is also positive for the latter (e.g. Piálek & Barton 1997, Faure et al 2008).<br />

Comparing different coalescence mo<strong>de</strong>ls already showed that the SAHH gene did not<br />

follow a neutral (M1) mo<strong>de</strong>l from both si<strong>de</strong>s of the EPR and was more compatible with a<br />

selective sweep mo<strong>de</strong>l (M3). Moreover, star-like topologies with the occurrence of<br />

potential surviving lineages associated with significantly negative Tajima and Fu & Li<br />

in<strong>de</strong>xes displayed on each part of the EPR strengthens the ‘selective sweep’ hypothesis for<br />

160


Chapitre 4<br />

SAHH. Even if the GlobX gene seems to coalesce neutrally, allele topologies were quite<br />

similar to those observed for the SAHH gene, and may also be interpreted as an even more<br />

recent sweep: a situation very difficult to <strong>de</strong>tect due to the lack of diversity (Berry et al<br />

1991). Occurrence of a selective sweep at loci SAHH and GlobX would be the best<br />

explanation to the lack (SAHH) or the very small (GlobX) differentiation between<br />

populations across the equatorial barrier, un<strong>de</strong>r the assumption of isolation in allopatry.<br />

The exten<strong>de</strong>d spread of advantageous alleles through the barrier will be therefore a<br />

convincing argument toward a recent change in the barrier porosity. The spread of an allele<br />

across a permeable barrier is however expected to be proportional to the selective<br />

coefficient of the advantage, suggesting that the locus GlobX could be more affected than<br />

the SAHH one (Morjan & Rieseberg 2004).<br />

Demographic expansion in the Southern EPR hypothesis<br />

A comparative phylogeographic approach across seven vent species using the<br />

mtCOI gene, including A. pompejana, revealed the occurrence of a concomitant<br />

<strong>de</strong>mographic expansion in Southern EPR in the last 0.5 My (Plouviez et al 2009). In the<br />

present study, Tajima’s D and Fu & Li in<strong>de</strong>xes were found significantly negative for all<br />

genes for the Southern cla<strong>de</strong> whereas it was significant at only one locus (SAHH) in the<br />

Northern one. If the selective sweep mo<strong>de</strong>l (M3) was favoured in the Southern cla<strong>de</strong>,<br />

indicating that at least one gene did not coalesce neutrally, the bottleneck mo<strong>de</strong>l (M2) was<br />

also significantly better than the neutral mo<strong>de</strong>l (M1). When consi<strong>de</strong>ring the M2 mo<strong>de</strong>l<br />

gene by gene, <strong>de</strong>parture to the neutral accumulation of mutations in polymorphism was<br />

significant for all genes in the Southern cla<strong>de</strong> whereas only the SAHH gene remained<br />

significant in the Northern cla<strong>de</strong>. Assuming the same distribution pattern of nucleotidic<br />

diversity across the studied genes (as expected for a <strong>de</strong>mographic event such as expansion,<br />

Hudson et al 1987, Galtier et al 2000), this multiloci approach is thus in accordance with<br />

the Southern population expansion hypothesized by Plouviez et al (2009). This result<br />

suggests that multiple extinction and re-colonisation events favouring the recurrence of<br />

bottlenecks through time (Haymon et al 1991, Tunnicliffe et al 1997, Jollivet et al 1999)<br />

are likely due to the higher extinction/instability rate observed in the Southern EPR when<br />

compared to the Northern one (Cormier 1997, Fouquet et al 1994).<br />

161


Chapitre 4<br />

The multilocus study performed on A. pompejana allowed to confirm that the<br />

equatorial barrier has played an important role in structuring populations of this species, as<br />

proposed by Plouviez et al (2009) in mtCOI. This geographic barrier was found semipermeable<br />

to gene flow with a hybrid zone that appears to be mainly ranged from 7°25’S<br />

to 14°S, with an asymmetric migration from North to South EPR and an introgression<br />

between mtCOI divergent lineages in the opposite sense. By i<strong>de</strong>ntifying one gene un<strong>de</strong>r<br />

selective sweep (SAHH) that crossed rapidly this barrier, and one putative balancing<br />

selected gene (PGM), this study thus constitutes a good example of different behaviours<br />

gene can display in hybrid zone. Comparison of allozymes, sequences and RFLP on PGM<br />

allowed to <strong>de</strong>tect an homoplasy of charge, and thus highlight that, even if strong (0.76%-<br />

mean divergence cla<strong>de</strong>s between North and South revealed by sequences), differentiation<br />

between populations could sometimes be masked in allozymes. In addition, discriminating<br />

a bottleneck mo<strong>de</strong>l from a selective sweep mo<strong>de</strong>l through the comparison of mtCOI and<br />

nuclear genes strongly support the Southern <strong>de</strong>mographic expansion hypothesis previously<br />

suggested by a multispecies analysis on mtCOI. This result highlight the necessity to use<br />

both multispecies and mutlilocus analyses to explore <strong>de</strong>mographic events.<br />

Acknowledgments<br />

We thank the chief scientists and ‘Nautile’ crews for their technical support and efforts<br />

during our oceanographic expeditions: PHARE2002, HOPE99 and BIOSPEEDO2004. We<br />

are very grateful to Stéphane Hour<strong>de</strong>z and Baptiste Faure for collecting and sorting<br />

alvinellid polychaetes. We are also truly in<strong>de</strong>bted to the sequencing genomic plateforms:<br />

GENOMER (<strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>, France) and GENOSCOPE (Evry, France) for<br />

the sequencing of the mtCOI fragment and our clone collection, respectively. This work<br />

was supported by the GDR Ecchis, the ANR-06-BDV-005 (Deep Oases: coord. D.<br />

Desbruyères) and ANR-05-BLAN-0407 (Alvi_Stress_Adapt). S. Plouviez was supported<br />

by a PhD grant from the Université Pierre et Marie Curie.<br />

162


Chapitre 4<br />

References<br />

Ban<strong>de</strong>lt HJ, Forster P, Rohl A (1999) Median-joining networks for inferring intraspecific<br />

phylogenies. Molecular Biology and Evolution 16, 37-48.<br />

Barton NH (1979) Gene Flow Past a Cline. Heredity 43, 333-339.<br />

Barton NH, Hewitt GM (1981) The Genetic-Basis of Hybrid Inviability in the Grasshopper<br />

Podisma-Pe<strong>de</strong>stris. Heredity 47, 367-383.<br />

Barton NH, Hewitt GM (1985) Analysis of Hybrid Zones. Annual Review of Ecology and<br />

Systematics 16, 113-148.<br />

Barton N, Bengtsson BO (1986) The Barrier to Genetic Exchange between Hybridizing<br />

Populations. Heredity 57, 357-376.<br />

Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and<br />

effective population sizes in n subpopulations by using a coalescent approach.<br />

Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of America 98,<br />

4563-4568.<br />

Berry AJ, Ajioka JW, Kreitman M (1991) Lack of Polymorphism on the Drosophila 4th<br />

Chromosome Resulting from Selection. Genetics 129, 1111-1117.<br />

Bierne N, Tanguy A, Faure M, et al. (2007) Mark-recapture cloning: a straightforward and<br />

cost-effective cloning method for population genetics of single copy nuclear DNA<br />

sequences in diploids. Molecular Ecology Notes.<br />

Bierne N, Borsa P, Daguin C, et al. (2003) Introgression patterns in the mosaic hybrid zone<br />

between Mytilus edulis and M-galloprovincialis. Molecular Ecology 12, 447-461.<br />

Birky CW, Maruyama T, Fuerst P (1983) An Approach to Population and Evolutionary<br />

Genetic Theory for Genes in Mitochondria and Chloroplasts, and Some Results.<br />

Genetics 103, 513-527.<br />

Birky CW, Fuerst P, Maruyama T (1989) Organelle Gene Diversity un<strong>de</strong>r Migration,<br />

Mutation, and Drift - Equilibrium Expectations, Approach to Equilibrium, Effects of<br />

Heteroplasmic Cells, and Comparison to Nuclear Genes. Genetics 121, 613-627.<br />

Black MB, Lutz RA, Vrijenhoek RC (1994) Gene Flow among Vestimentiferan Tube<br />

Worm (Riftia-Pachyptila) Populations from Hydrothermal Vents of the Eastern<br />

Pacific. Marine Biology 120, 33-39.<br />

Boon E, Faure MF, Bierne N (2009) The Flow of Antimicrobial Pepti<strong>de</strong> Genes Through a<br />

Genetic Barrier Between Mytilus edulis and M-galloprovincialis. Journal of<br />

Molecular Evolution 68, 461-474.<br />

Broughton RE, Harrison RG (2003) Nuclear gene genealogies reveal historical,<br />

<strong>de</strong>mographic and selective factors associated with speciation in field crickets. Genetics<br />

163, 1389-1401.<br />

Burton RS, Lee BN (1994) Nuclear and Mitochondrial Gene Genealogies and Allozyme<br />

Polymorphism across a Major Phylogeographic Break in the Copepod Tigriopus-<br />

Californicus. Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of<br />

America 91, 5197-5201.<br />

Castric V, Vekemans X (2007) Evolution un<strong>de</strong>r strong balancing selection: how many<br />

codons <strong>de</strong>termine specificity at the female self-incompatibility gene SRK in<br />

Brassicaceae? Bmc Evolutionary Biology 7.<br />

Castric V, Bechsgaard J, Schierup MH, Vekemans X (2008) Repeated Adaptive<br />

Introgression at a Gene un<strong>de</strong>r Multiallelic Balancing Selection. Plos Genetics 4.<br />

Craddock C, Hoeh WR, Lutz RA, Vrijenhoek RC (1995) Extensive Gene Flow among<br />

Mytilid (Bathymodiolus-Thermophilus) Populations from Hydrothermal Vents of the<br />

Eastern Pacific. Marine Biology 124, 137-146.<br />

163


Chapitre 4<br />

Chevaldonné P, Jollivet D, Vangriesheim A, Desbruyeres D (1997) Hydrothermal-vent<br />

alvinellid polychaete dispersal in the eastern Pacific .1. Influence of vent site<br />

distribution, bottom currents, and biological patterns. Limnology and Oceanography<br />

42, 67-80.<br />

Chevaldonné P, Jollivet D, Desbruyeres D, Lutz RA, Vrijenhoek RC (2002) Sister-species<br />

of eastern Pacific hydrothermal vent worms (Ampharetidae, Alvinellidae,<br />

Vestimentifera) provi<strong>de</strong> new mitochondrial COI clock calibration. Cahiers De<br />

Biologie Marine 43, 367-370.<br />

Cormier M-H (1997) The ultrafast East Pacific Rise: instability of the plate boundary and<br />

implications for accretionary processes. Philosophical Transactions of the Royal<br />

Society of London 355, 341-367.<br />

Cunningham CW, Collins TM (1998) Beyond area relationships : extinction and<br />

recolonization in molecular marine biogeography. In : Molecular Approaches to<br />

Ecology and Evolution. DeSalle R, Schierwater B (eds), Birkhäser Verlag Basel, 297-<br />

321.<br />

Excoffier L, Laval G, Schnei<strong>de</strong>r S (2005) Arlequin (version 3.0): An integrated software<br />

package for population genetics data analysis. Evolutionary Bioinformatics 1, 47-50.<br />

Faure B, Jollivet D, Tanguy A, Bonhomme F, Bierne N (2009) Speciation in the <strong>de</strong>ep sea:<br />

multi-locus analysis of divergence and gene flow between two hybridizing species of<br />

hydrothermal vent mussels. PLoS One 4, e6485.<br />

Faure MF, David P, Bonhomme F, Bierne N (2008) Genetic hitchhiking in a subdivi<strong>de</strong>d<br />

population of Mytilus edulis. Bmc Evolutionary Biology 8.<br />

Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification<br />

of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates.<br />

Molecular Marine Biology and Biotechnology 3, 294-299.<br />

Fouquet Y, Auzen<strong>de</strong> JM, Ballu V, et al. (1994) Variability of the Present-Day<br />

Hydrothermal Activity Along an Ultra-Fast Spreading Ridge - East Pacific Rise<br />

between 17-Degrees-C and 19-Degrees-S (Naudur Cruise). Comptes Rendus De L<br />

Aca<strong>de</strong>mie Des Sciences Serie Ii 319, 1399-1406.<br />

Francheteau J, Armijo R, Cheminee JL, et al. (1990) 1 Ma East Pacific Rise Oceanic-Crust<br />

and Uppermost Mantle Exposed by Rifting in Hess Deep (Equatorial Pacific-Ocean).<br />

Earth and Planetary Science Letters 101, 281-295.<br />

Gagnière N, Jollivet D, Boutet I, et al (submitted) Insights metazoan evolution from<br />

Alvinella pompejana cDNAs. Genome Research.<br />

Galtier N, Depaulis F, Barton NH (2000) Detecting bottlenecks and selective sweeps from<br />

DNA sequence polymorphism. Genetics 155, 981-987.<br />

Garnier-Géré P, Dillmann C (1992) A Computer-Program for Testing Pairwise Linkage<br />

Disequilibria in Subdivi<strong>de</strong>d Populations. Journal of Heredity 83, 239-239.<br />

Glémin S, Gau<strong>de</strong> T, Guillemin ML, et al. (2005) Balancing selection in the wild: Testing<br />

population genetics theory of self-incompatibility in the rare species Brassica insularis.<br />

Genetics 171, 279-289.<br />

Godinho R, Crespo EG, Ferrand N (2008) The limits of mtDNA phylogeography: complex<br />

patterns of population history in a highly structured Iberian lizard are only revealed by<br />

the use of nuclear markers. Molecular Ecology 17, 4670-4683.<br />

Graham JH, Freeman DC, McArthur ED (1995) Narrow Hybrid Zone between 2<br />

Subspecies of Big Sagebrush (Artemisia-Tri<strong>de</strong>ntata, Asteraceae) .2. Selection<br />

Gradients and Hybrid Fitness. American Journal of Botany 82, 709-716.<br />

Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis<br />

program for Windows 95/98/NT. Nucleic Acids Symposium Series 41, 95-98.<br />

164


Chapitre 4<br />

Harrison S, Hastings A (1996) Genetic and evolutionary consequences of metapopulation<br />

structure. Trends in Ecology & Evolution 11, 180-183.<br />

Hasegawa M, Kishino H, Yano T-A (1985) Dating of the human-ape splitting by a<br />

molecular clock of mitochondrial DNA. Journal of Molecular Evolution 22, 160-174.<br />

Haymon RM, Fornari DJ, Edwards MH, et al. (1991) Hydrothermal Vent Distribution<br />

Along the East Pacific Rise Crest (9-Degrees-09'-54'n) and Its Relationship to<br />

Magmatic and Tectonic Processes on Fast-Spreading Midocean Ridges. Earth and<br />

Planetary Science Letters 104, 513-534.<br />

Hey R, Baker E, Bohnenstiehl D, et al. (2004) Tectonic/volcanic segmentation and<br />

controls on hydrothermal venting along Earth's fastest seafloor spreading system, EPR<br />

27 <strong>de</strong>grees-32 <strong>de</strong>grees S. Geochemistry Geophysics Geosystems 5.<br />

Hey J, Nielsen R (2007) Integration within the Felsenstein equation for improved Markov<br />

chain Monte Carlo methods in population genetics. Proceedings of the National<br />

Aca<strong>de</strong>my of Sciences of the United States of America 104, 2785-2790.<br />

Hudson RR, Kaplan NL (1985) Statistical Properties of the Number of Recombination<br />

Events in the History of a Sample of DNA-Sequences. Genetics 111, 147-164.<br />

Hudson RR, Kreitman M, Agua<strong>de</strong> M (1987) A Test of Neutral Molecular Evolution Based<br />

on Nucleoti<strong>de</strong> Data. Genetics 116, 153-159.<br />

Hudson RR, Turelli M (2003) Stochasticity overrules the "three-times rule": Genetic drift,<br />

genetic draft, and coalescence times for nuclear loci versus mitochondrial DNA.<br />

Evolution 57, 182-190.<br />

Hurtado L, Lutz R, Vrijenhoek R (2004) Distinct patterns of genetic differentiation among<br />

annelids of eastern Pacific hydrothermal vents. Molecular Ecology 13, 2603-2615.<br />

Irwin DE (2002) Phylogeographic breaks without geographic barriers to gene flow.<br />

Evolution 56, 2383-2394.<br />

Johnson SB, Young CR, Jones WJ, Waren A, Vrijenhoek RC (2006) Migration, isolation,<br />

and speciation of hydrothermal vent limpets (Gastropoda; Lepetodrilidae) across the<br />

Blanco transform fault Biological Bulletin 210, 140-157.<br />

Jollivet D, Desbruyeres D, Bonhomme F, Moraga D (1995) Genetic Differentiation of<br />

Deep-Sea Hydrothermal Vent Alvinellid Populations (Annelida, Polychaeta) Along<br />

the East Pacific Rise. Heredity 74, 376-391.<br />

Jollivet D (1996) Specific and genetic diversity at <strong>de</strong>ep-sea hydrothermal vents: An<br />

overview. Biodiversity and Conservation 5, 1619-1653.<br />

Jollivet D, Chevaldonne P, Planque B (1999) Hydrothermal-vent alvinellid polychaete<br />

dispersal in the eastern Pacific. 2. A metapopulation mo<strong>de</strong>l based on habitat shifts.<br />

Evolution 53, 1128-1142.<br />

Keenan CP (1991) Phylogeny of Australian Species of Flatheads (Teleostei,<br />

Platycephalidae) as Determined by Allozyme Electrophoresis. Journal of Fish Biology<br />

39, 237-249.<br />

Kimura M (1953) "Stepping stone" mo<strong>de</strong>l of population. Annual Report Nationnal<br />

Institute of Genetics. Japan 3, 62-63.<br />

Kimura M, Weiss GH (1964) Stepping Stone Mo<strong>de</strong>l of Population Structure + Decrease of<br />

Genetic Correlation with Distance. Genetics 49, 561<br />

Kruuk LEB, Baird SJE, Gale KS, Barton NH (1999) A comparison of multilocus clines<br />

maintained by environmental adaptation or by selection against hybrids. Genetics 153,<br />

1959-1971.<br />

Kureth CL, Rea DK (1981) Large-Scale Oblique Features in an Active Transform-Fault,<br />

the Wilkes Fracture-Zone near 9-Degrees-S on the East Pacific Rise. Marine<br />

Geophysical Researches 5, 119-137.<br />

165


Chapitre 4<br />

Lemaire C, Versini JJ, Bonhomme F (2005) Maintenance of genetic differentiation across<br />

a transition zone in the sea: discordance between nuclear and cytoplasmic markers.<br />

Journal of Evolutionary Biology 18, 70-80.<br />

Lewontin RC, Krakauer J (1973) Distribution of Gene Frequency as a Test of Theory of<br />

Selective Neutrality of Polymorphisms. Genetics 74, 175-195.<br />

Lutz RA, Fritz LW, Cerrato RM (1988) A Comparison of Bivalve (Calyptogena-<br />

Magnifica) Growth at 2 Deep-Sea Hydrothermal Vents in the Eastern Pacific. Deep-<br />

Sea Research Part a-Oceanographic Research Papers 35, 1793<br />

Malécot G (1950) Quelques schémas probabilistes sur la variabilité <strong>de</strong>s populations<br />

naturelles. Annales <strong>de</strong> l’Université <strong>de</strong> Lyon Sciences 13, 37-60.<br />

Mantel N (1967) The <strong>de</strong>tection of disease clustering and a generalized regression<br />

approach. Cancer Research 27, 209-220.<br />

Martinsen GD, Whitham TG, Turek RJ, Keim P (2001) Hybrid populations selectively<br />

filter gene introgression between species. Evolution 55, 1325-1335.<br />

Morjan CL, Rieseberg LH (2004) How species evolve collectively: implications of gene<br />

flow and selection for the spread of advantageous alleles. Molecular Ecology 13,<br />

1341-1356.<br />

Muirhead CA (2001) Consequences of population structure on genes un<strong>de</strong>r balancing<br />

selection. Evolution 55, 1532-1541.<br />

Naar DF, Hey RN (1989) Speed Limit for Oceanic Transform Faults. Geology 17, 420-<br />

422.<br />

Nagylaki T (1975) Conditions for Existence of Clines. Genetics 80, 595-615.<br />

Neigel JE, Avise JC (1993) Application of a Random-Walk Mo<strong>de</strong>l to Geographic<br />

Distributions of Animal Mitochondrial-DNA Variation. Genetics 135, 1209-1220.<br />

Nosil P, Harmon LJ, Seehausen O (2009) Ecological explanations for (incomplete)<br />

speciation. Trends in Ecology & Evolution 24, 145-156.<br />

OhUigin C, Sato A, Klein J (1997) Evi<strong>de</strong>nce for convergent evolution of A and B blood<br />

group antigens in primates. Human Genetics 101, 141-148.<br />

Pannell JR, Charlesworth B (2000) Effects of metapopulation process on measures of<br />

genetic diversity. Philosophical Transactions of the Royal Society of London 355,<br />

1851-1864.<br />

Pasteur N, Pasteur G, Bonhomme F, Catalan J, Britton-Davidian J (1987) Manuel<br />

technique <strong>de</strong> génétique par électrophorèse <strong>de</strong> protéines. Paris: Lavoisier.<br />

Piálek J, Barton NH (1997) The spread of an advantageous allele across a barrier: The<br />

effects of random drift and selection against heterozygotes. Genetics 145, 493-504.<br />

Piccino P, Viard F, Sarradin PM, et al. (2004) Thermal selection of PGM allozymes in<br />

newly foun<strong>de</strong>d populations of the thermotolerant vent polychaete Alvinella<br />

pompejana. Proceedings of the Royal Society of London Series B-Biological Sciences<br />

271, 2351-2359.<br />

Plouviez S, Shank TM, Faure B, et al. (2009) Comparative phylogeography among<br />

hydrothermal vent species along the East Pacific Rise reveals vicariant processes and<br />

population expansion in the South. Molecular Ecology 18, 3903-3917.<br />

Pogson GH, Mesa KA, Boutilier RG (1995) Genetic Population-Structure and Gene Flow<br />

in the Atlantic Cod Gadus-Morhua - a Comparison of Allozyme and Nuclear Rflp<br />

Loci. Genetics 139, 375-385.<br />

Pradillon F, Shillito B, Young CM, Gaill F (2001) Deep-seaecology - Developmental<br />

arrest in vent worm embryos. Nature 413, 698-699.<br />

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using<br />

multilocus genotype data. Genetics 155, 945-959.<br />

166


Chapitre 4<br />

Reid JL (1997) On the total geostrophic circulation of the South-Pacific ocean: flow<br />

patterns, tracers, and transports. Progress in Oceanography 39, 263-352.<br />

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA<br />

polymorphism analyses by the coalescent and other methods. Bioinformatics 19, 2496-<br />

2497.<br />

Schierup MH, Charlesworth D, Vekemans X (2000a) The effect of hitch-hiking on genes<br />

linked to a balanced polymorphism in a subdivi<strong>de</strong>d population. Genetical Research<br />

76, 63-73.<br />

Schierup MH, Vekemans X, Charlesworth D (2000b) The effect of subdivision on<br />

variation at multi-allelic loci un<strong>de</strong>r balancing selection. Genetical Research 76, 51-62.<br />

Schlager W (2008) Earth’s layers, their cycles and Earth system science. Austrian Journal<br />

of Earth Sciences 101, 4-16.<br />

Slatkin M (1977) Gene flow and genetic drift in a species subject to frequent local<br />

extinction. Theoretical Population Biology 12, 253-262.<br />

Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology &<br />

Evolution 15, 199-203.<br />

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular evolutionary genetics<br />

analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24, 1596-<br />

1599.<br />

Templeton AR (1981) Mechanisms of Speciation - a Population Genetic Approach. Annual<br />

Review of Ecology and Systematics 12, 23-48.<br />

Thompson JD, Higgins DG, Gibson TJ (1994) Clustal-W - Improving the Sensitivity of<br />

Progressive Multiple Sequence Alignment through Sequence Weighting, Position-<br />

Specific Gap Penalties and Weight Matrix Choice. Nucleic Acids Research 22, 4673-<br />

4680.<br />

Ting CT, Tsaur SC, Wu CI (2000) The phylogeny of closely related species as revealed by<br />

the genealogy of a speciation gene, Odysseus. Proceedings of the National Aca<strong>de</strong>my<br />

of Sciences of the United States of America 97, 5313-5316.<br />

Tunnicliffe V, Embley RW, Hol<strong>de</strong>n JF, et al. (1997) Biological colonization of new<br />

hydrothermal vents following an eruption on Juan <strong>de</strong> Fuca Ridge. Deep-Sea Research<br />

Part I-Oceanographic Research Papers 44, 1627.<br />

Vrijenhoek RC (1997) Gene flow and genetic diversity in naturally fragmented<br />

metapopulations of <strong>de</strong>ep-sea hydrothermal vent animals. Journal of Heredity 88, 285-<br />

293.<br />

Weir BS, Cockerham CC (1984) Estimating F-Statistics for the Analysis of Population-<br />

Structure. Evolution 38, 1358-1370.<br />

Whitlock MC, McCauley DE (1990) Some population genetic consequences of colony<br />

formation and extinction: genetic correlations within foundings groups. Evolution 44,<br />

1717-1724.<br />

Wood TE, Burke JM, Rieseberg LH (2005) Parallel genotypic adaptation: when evolution<br />

repeats itself. Genetica 123, 157-170.<br />

Wright S (1931) Evolution in men<strong>de</strong>lian populations. Genetics 16, 97-159.<br />

Wright S (1937) The distribution of gene frequencies in populations. Proceedings of the<br />

National Aca<strong>de</strong>my of Sciences of the United States of America 23, 307-320.<br />

Wright S (1943) Isolation by distance. Genetics 28, 114-138.<br />

Zhang JZ, Kumar S (1997) Detection of convergent and parallel evolution at the amino<br />

acid sequence level. Molecular Biology and Evolution 14, 527-536.<br />

167


Chapitre 4<br />

Supplementary material<br />

Locus 13°N 7°25'S 14°S 17°25'S 17°34'S 18°33'S 21°25'-33'S<br />

PGM<br />

(N) 40 17 15 34 19 52 35<br />

78 0.0375 0 0 0 0 0 0<br />

90 0.3875 0.5588 0.7333 0.6912 0.8684 0.8558 0.7429<br />

100 0.5 0.2353 0.2333 0.2941 0.1053 0.1442 0.2429<br />

112 0.075 0.1765 0.0333 0.0147 0.0263 0 0.0143<br />

120 0 0.0294 0 0 0 0 0<br />

H exp. 0.5928 0.6003 0.4067 0.4356 0.2341 0.2469 0.389<br />

H n.b. 0.6003 0.6185 0.4207 0.4421 0.2404 0.2493 0.3946<br />

H obs. 0.575 0.5294 0.2 0.3529 0.2632 0.25 0.2571<br />

Fis 0.018 0.168 0.533* 0.213 -0.098 -0.003 0.352*<br />

MPI<br />

(N) 40 17 15 34 19 52 35<br />

70 0.0125 0 0 0 0 0 0<br />

92 0.0375 0 0 0 0.0263 0.0096 0<br />

96 0.35 0.3235 0.1333 0.0882 0.1579 0.1058 0.0714<br />

100 0.6 0.6765 0.8667 0.9118 0.8158 0.8846 0.9286<br />

H exp. 0.5159 0.4377 0.2311 0.1609 0.3089 0.2062 0.1327<br />

H n.b. 0.5225 0.451 0.2391 0.1633 0.3172 0.2082 0.1346<br />

H obs. 0.675 0.2941 0.1333 0.1765 0.3684 0.1923 0.0857<br />

Fis -0.308 0.370* 0.451* -0.097 -0.167 0.077 0.366<br />

GPI<br />

(N) 40 17 15 34 19 52 35<br />

70 0.0125 0 0 0 0 0 0<br />

100 0.9875 1 1 0.9118 0.9737 0.9904 0.9429<br />

120 0 0 0 0.0882 0.0263 0.0096 0.0571<br />

H exp. 0.0247 0 0 0.1609 0.0512 0.019 0.1078<br />

H n.b. 0.025 0 0 0.1633 0.0526 0.0192 0.1093<br />

H obs. 0.025 0 0 0.0588 0.0526 0.0192 0.0571<br />

Fis -0.012 - - 0.644* 0.000 0.000 0.481<br />

6PGD<br />

(N) 40 17 15 34 19 52 35<br />

80 0 0 0 0 0 0.0096 0<br />

90 0.2 0.4118 0.2333 0.2206 0.3158 0.2019 0.1571<br />

100 0.775 0.5588 0.7667 0.75 0.6842 0.6154 0.8286<br />

110 0.025 0.0294 0 0.0294 0 0.1731 0.0143<br />

H exp. 0.3588 0.5173 0.3578 0.388 0.4321 0.5505 0.2886<br />

H n.b. 0.3633 0.533 0.3701 0.3938 0.4438 0.5558 0.2928<br />

H obs. 0.3 0.4118 0.2 0.3235 0.5263 0.4808 0.2286<br />

Fis 0.190 0.261 0.468 0.187 -0.192 0.136 0.222<br />

168


IDHhaut<br />

(N) 40 17 15 34 19 52 35<br />

100 0.9 0.8824 0.8333 0.9853 0.8947 0.9712 0.9429<br />

120 0.0625 0.1176 0.1333 0 0.1053 0.0096 0.0571<br />

140 0.0375 0 0.0333 0.0147 0 0.0192 0<br />

H exp. 0.1847 0.2076 0.2867 0.029 0.1884 0.0564 0.1078<br />

H n.b. 0.187 0.2139 0.2966 0.0294 0.1935 0.0569 0.1093<br />

H obs. 0.15 0.2353 0.3333 0.0294 0.2105 0.0577 0.1143<br />

Fis 0.204 -0.097 -0.129 0.000 -0.091 -0.013 -0.046<br />

ACP<br />

(N) 40 17 15 34 19 52 35<br />

90 0.0125 0 0 0 0.0263 0.0096 0<br />

100 0.9875 0.6765 0.9667 1 0.9737 0.9808 0.8143<br />

120 0 0.3235 0.0333 0 0 0.0096 0.1857<br />

H exp. 0.0247 0.4377 0.0644 0 0.0512 0.0379 0.3024<br />

H n.b. 0.025 0.451 0.0667 0 0.0526 0.0383 0.3068<br />

H obs. 0.025 0.5294 0.0667 0 0.0526 0.0385 0.3714<br />

Fis 0.000 -0.150 0.000 - 0.000 -0.005 -0.214<br />

LAP<br />

(N) 40 17 15 34 19 52 35<br />

90 0 0 0 0.0294 0 0 0<br />

100 1 1 0.9333 0.9706 1 1 1<br />

110 0 0 0.0667 0 0 0 0<br />

H exp. 0 0 0.1244 0.0571 0 0 0<br />

H n.b. 0 0 0.1287 0.0579 0 0 0<br />

H obs. 0 0 0.1333 0.0588 0 0 0<br />

Fis - - -0.037 -0.015 - - -<br />

TPI<br />

(N) 40 17 15 34 19 52 35<br />

82 0 0.0294 0 0.0294 0 0.0481 0.0143<br />

90 0 0.0294 0 0.1471 0.0263 0.3269 0.0571<br />

100 0.3875 0.8529 0.9333 0.6912 0.6842 0.6154 0.6857<br />

110 0.5875 0.0882 0.0667 0.1324 0.2895 0.0096 0.2429<br />

120 0.025 0 0 0 0 0 0<br />

H exp. 0.5041 0.263 0.1244 0.4823 0.4474 0.512 0.4673<br />

H n.b. 0.5104 0.2709 0.1287 0.4895 0.4595 0.517 0.4741<br />

H obs. 0.6 0.1765 0 0.6176 0.4211 0.75 0.4<br />

Fis -0.232 0.358 1.000* -0.257 0.086 -0.457 0.158<br />

HK<br />

(N) 40 17 15 34 19 52 35<br />

96 0.025 0 0 0.0441 0.0263 0.0096 0<br />

100 0.5625 0.8824 0.6667 0.6618 0.7895 0.7308 0.7571<br />

104 0.4125 0.1176 0.3333 0.2941 0.1842 0.2596 0.2429<br />

H exp. 0.5128 0.2076 0.4444 0.4736 0.3421 0.3985 0.3678<br />

H n.b. 0.5193 0.2139 0.4598 0.4807 0.3514 0.4024 0.3731<br />

H obs. 0.725 0.2353 0.2667 0.2059 0.3158 0.4231 0.4857<br />

Fis -0.388 -0.097 0.429 0.580*** 0.104 -0.052 -0.308<br />

Chapitre 4<br />

169


Chapitre 4<br />

4.2. B. thermophilus<br />

4.2.1. Résumé <strong>de</strong>s résultats<br />

Contrairement à l’espèce A. pompejana qui associe divergence sur le mtCOI et<br />

différenciation géographique <strong>de</strong> part et d’autre <strong>de</strong> la barrière équatoriale (constituée en<br />

partie par le complexe <strong>de</strong> failles transformantes Discovery/Gofar), la moule hydrothermale<br />

B. thermophilus présente <strong>de</strong>ux lignées divergentes sur le gène mitochondrial sans situation<br />

d’allopatrie stricte <strong>de</strong> 13°N à 7°25’S, ce qui rend le positionnement <strong>de</strong> la barrière entre<br />

populations Nord et Sud EPR difficile sur la base d’un seul gène. De plus, les 2 lignées<br />

mitochondriales étant mélangées sur les sites 13°N, 9°50N et 7°25S, plusieurs scenarii <strong>de</strong><br />

colonisation <strong>pour</strong>raient expliquer la distribution géographique <strong>de</strong> ces lignées :<br />

(1) un isolement physique puis génétique <strong>de</strong> part et d’autre <strong>de</strong> la barrière par<br />

vicariance, suivi d’une remise en contact secondaire au Nord <strong>de</strong> l’EPR liée aux<br />

caractéristiques particulières <strong>de</strong> dispersion chez cette espèce,<br />

(2) un isolement sympatrique (parapatrique) <strong>de</strong>s <strong>de</strong>ux lignées mitochondriales au Nord<br />

suivi <strong>de</strong> la migration d’une seule <strong>de</strong> ces lignées vers le Sud à travers la barrière,<br />

(3) l’extinction <strong>de</strong> la lignée nord au Sud <strong>de</strong> l’EPR due au taux <strong>de</strong> remaniement<br />

tectonique <strong>de</strong>s sites plus élevé dans cette région.<br />

Une comparaison <strong>de</strong>s flux <strong>de</strong> gènes entre le gène mtCOI, <strong>de</strong>s marqueurs allozymiques et le<br />

polymorphisme <strong>de</strong> séquences <strong>de</strong> quatre gènes nucléaires a ainsi été réalisée afin <strong>de</strong> :<br />

(1) préciser la position géographique <strong>de</strong> la barrière aux flux <strong>de</strong> gènes<br />

(2) déterminer le <strong>de</strong>gré <strong>de</strong> perméabilité <strong>de</strong> cette barrière, en évaluant la migration à<br />

travers celle-ci (intensité, asymétrie)<br />

(3) quantifier l’hybridation potentielle entre les 2 lignées mitochondriales<br />

(4) vérifier par l’approche multi-locus l’hypothèse d’expansion démographique au Sud<br />

<strong>de</strong> la dorsale proposée par Plouviez et al (2009).<br />

L’analyse <strong>de</strong>s allozymes effectuée <strong>de</strong> 7°25’S à 21°33’S n’a pas permis <strong>de</strong> mettre en<br />

évi<strong>de</strong>nce <strong>de</strong> structure génétique sur cette portion <strong>de</strong> dorsale et confirme l’absence<br />

d’isolement par la distance chez B. thermophilus suggérée par Craddock et al (1995) sur<br />

l’EPR Nord et les sites Galápagos. Les quatre gènes nucléaires présentent, quant à eux, une<br />

170


Chapitre 4<br />

divergence marquée en 2 cla<strong>de</strong>s d’allèles (Fig. 4.14) associée à une différenciation<br />

génétique entre les champs hydrothermaux 7°25’S et 14°S, et confirment donc la<br />

localisation géographique supposée d’une barrière dans cette zone géographique. Les<br />

gènes S-A<strong>de</strong>nosyl Homocystéine Hydrolase (SAHH) et Sulfotransférase 1 (Sulfo 1)<br />

présentent ainsi <strong>de</strong>ux cla<strong>de</strong>s divergents répartis respectivement <strong>de</strong> part et d’autre <strong>de</strong> cette<br />

barrière géographique. De plus, si les cla<strong>de</strong>s divergents observés sur le gène Lysozyme<br />

(Lyso) se distribuent sur l’ensemble <strong>de</strong> la dorsale, un in<strong>de</strong>l <strong>de</strong> 1 pb présent dans les <strong>de</strong>ux<br />

cla<strong>de</strong>s (et donc impliquant une recombinaison entre ces cla<strong>de</strong>s) est structuré<br />

géographiquement. En effet, la « délétion » est trouvée uniquement au Sud <strong>de</strong> la barrière,<br />

indiquant que celle-ci serait récemment imperméable au passage <strong>de</strong>s nouvelles mutations,<br />

au moins <strong>pour</strong> ce gène. La migration détectée à travers cette barrière via le logiciel<br />

Migrate-n relativement faible serait alors probablement issue d’un contact secondaire<br />

ancien. Si cette migration est asymétrique <strong>pour</strong> l’ensemble <strong>de</strong>s gènes, le sens <strong>de</strong> migration<br />

dépend du gène considéré (Nord vers Sud <strong>pour</strong> SAHH et Lyso, Sud vers Nord <strong>pour</strong><br />

mtCOI, Sulfo 1 et Sulfo 2), indiquant que <strong>de</strong>s forces <strong>de</strong> sélection (e.g. sélection contre les<br />

hybri<strong>de</strong>s) <strong>pour</strong>raient agir différemment sur ces gènes entre le Nord et le Sud <strong>de</strong> la barrière<br />

selon le ‘background’ génétique <strong>de</strong> chaque lignée.<br />

171


Chapitre 4<br />

0,002<br />

0,001<br />

SAHH<br />

Lyso<br />

0,002<br />

0,001<br />

Sulfo 1 Sulfo 2<br />

Fig. 4.14 Arbres phylogénétiques réalisés selon la métho<strong>de</strong>s <strong>de</strong>s plus proches voisins<br />

sur les gènes nucléaires <strong>de</strong> B. thermophilus à partir <strong>de</strong>s jeux <strong>de</strong> séquences <strong>de</strong> qualité 2<br />

(2 allèles maximum par individu).<br />

En bleu et rouge : séquences d’individus échantillonnées respectivement au Nord<br />

(9°50’N-7°25’S) et au Sud (14°S-21°33’S) <strong>de</strong> la barrière.<br />

172


Chapitre 4<br />

L’utilisation d’un modèle d’ « Isolement avec Migration » (IMa, Hey & Nielsen<br />

2007) suggère une introgression d’allèles récente <strong>pour</strong> tous les gènes nucléaires à<br />

l’exception du gène Lyso (introgression ancienne et absence d’introgression récente<br />

apparente) principalement orientée <strong>de</strong> la lignée mtCOI majoritaire au Nord vers celle<br />

majoritairement échantillonnée au Sud (Fig. 4.15). À l’exception du gène Lyso, le sens <strong>de</strong><br />

l’introgression <strong>de</strong>s gènes s’oppose au sens <strong>de</strong> migration détecté par Migrate_n, renforçant<br />

l’hypothèse que la sélection contre les hybri<strong>de</strong>s produits freine la migration <strong>de</strong>s allèles à<br />

travers la barrière. Un balayage sélectif (ou effet autostop) détecté sur le gène Lyso via le<br />

logiciel Sweep_bott et conforté par les D <strong>de</strong> Tajima et F* <strong>de</strong> Fu & Li significativement<br />

négatifs au nord et au sud. L’hypothèse d’un passage d’allèle avantageux à travers la<br />

barrière <strong>pour</strong>rait expliquer que le sens d’introgression soit i<strong>de</strong>ntique au sens <strong>de</strong> migration<br />

sur ce gène (balayage d’un allèle avantageux issu du Nord <strong>de</strong> l’EPR vers le Sud). Si les<br />

gènes Sulfo 1 et 2 contribuent exclusivement à la significativité du test HKA, ces 2 locus<br />

ne présentent pas <strong>de</strong> coalescents s’écartant <strong>de</strong> l’attendu neutre (Sweep_bott) et l’absence<br />

<strong>de</strong> mutation fixée entre B. thermophilus et l’espèce Atlantique B. azoricus suggère que ces<br />

gènes sont plutôt soumis à une forte sélection purifiante.<br />

Les indices <strong>de</strong> Tajima et Fu & Li significatifs au Sud <strong>pour</strong> l’ensemble <strong>de</strong>s gènes ne le<br />

sont que <strong>pour</strong> le gène Lyso au nord. Ces résultats complémentés par les analyses<br />

d’adéquation à un modèle <strong>de</strong> goulot d’étranglement (bottleneck) effectuées par le logiciel<br />

Sweep_bott et la mise en évi<strong>de</strong>nce par le logiciel IMa d’une taille efficace <strong>de</strong> la lignée<br />

mitochondriale Sud supérieure à celles <strong>de</strong>s lignée Nord et la lignée ancestrale (Fig. 4.15),<br />

confirment l’hypothèse d’expansion démographique au Sud émise par l’approche multiespèces<br />

sur le gène mtCOI (Plouviez et al 2009).<br />

L’approche multi-locus effectuée sur B. thermophilus a donc permis <strong>de</strong> préciser la<br />

position géographique <strong>de</strong> la barrière aux flux <strong>de</strong> gène (7°25’S-14°S), <strong>de</strong> confirmer sa date<br />

<strong>de</strong> mise en place (environ 1,3 Ma, Fig. 4.15) proposée par l’approche multispécifique et, <strong>de</strong><br />

montrer la faible perméabilité <strong>de</strong> celle-ci aux flux <strong>de</strong> gènes. De plus, cette approche a<br />

permis <strong>de</strong> révéler qu’une introgression d’allèles entre les lignées mtCOI, majoritairement<br />

orientée <strong>de</strong> la lignée Nord vers la lignée Sud <strong>pour</strong>rait expliquer <strong>de</strong>s flux migratoires<br />

principalement orientés vers le nord (introgression préférentielle dans l’espèce migrante) :<br />

une situation relativement pertinente si l’on considère que les populations situées au nord<br />

<strong>de</strong> 9°50N sont en limite d’aire géographique. De plus, un certain nombre <strong>de</strong> gènes <strong>pour</strong>rait<br />

173


Chapitre 4<br />

avoir été potentiellement soumis à <strong>de</strong>s forces <strong>de</strong> sélection homogénéisante (balayage<br />

sélectif : Lyso, sélection purifiante probable : Sulfo 1 et 2). L’expansion démographique au<br />

Sud <strong>de</strong> la dorsale suggérée par l’approche multispécifique (Plouviez et al 2009) et<br />

confirmée sur le polychète A. pompejana par l’utilisation <strong>de</strong> gènes nucléaires, est<br />

également vérifiée sur les gènes nucléaires <strong>de</strong> B. thermophilus.<br />

A<br />

B<br />

C<br />

174<br />

Fig. 4.15 Distributions <strong>de</strong>s probabilités postérieures marginales du temps <strong>de</strong><br />

divergence entre lignées (A), <strong>de</strong> la taille efficace <strong>de</strong> la population ancestrale et <strong>de</strong> ces<br />

populations filles, (B), et <strong>de</strong>s taux <strong>de</strong> flux <strong>de</strong> gènes directionnels entre lignées (C),<br />

estimées par un modèle d’Isolement avec Migration via IMa.


Chapitre 4<br />

4.2.2. Article en préparation<br />

Polymorphism to divergence analysis of nuclear genes reveals a semi-permeable<br />

transition zone between the Northern and Southern East Pacific Rise in the <strong>de</strong>ep-sea<br />

mussel Bathymodiolus thermophilus<br />

Plouviez S. 1,2 , Faure B. 3 , Le Guen D. 1,2 , Lallier F.H. 1,2 , Bierne N. 4 , Jollivet D. 1,2<br />

1 Université Pierre et Marie Curie-Paris 6, Laboratoire Adaptation et Diversité en Milieu<br />

Marin, <strong>Roscoff</strong>, France<br />

2 CNRS, UMR 7144 <strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>, BP 74, Place Georges Teissier, 29682,<br />

<strong>Roscoff</strong>, France<br />

3 Present address: Department of Biology, Pennsylvania State University, University Park,<br />

Pennsylvania 16802, USA<br />

4 Département Biologie Intégrative, Institut <strong>de</strong>s Sciences <strong>de</strong> l’Evolution, UMR 5554,<br />

Université Montpellier 2, Sète, France<br />

Key words: allozymes, nuclear genes, divergence, coalescence, barrier, secondary contact<br />

zone<br />

175


Chapitre 4<br />

Abstract<br />

Comparative phylogeography of <strong>de</strong>ep-sea hydrothermal vent species previously showed<br />

the occurrence of a genetic break between the Northern and Southern parts of the East<br />

Pacific Rise (EPR) that gave birth to a more-or-less sharp transition zone <strong>de</strong>pending on<br />

species. If some species, such as the polychaete Alvinella pompejana, displays a clear<br />

geographic separation between two mitochondrial lineages, position of the barrier was<br />

more difficult to establish for the <strong>de</strong>ep-sea vent mussel Bathymodiolus thermophilus<br />

because of a smooth clinal distribution of the mt lineages along the North EPR. The<br />

present study aims at comparing geographic distribution of the mtCOI lineages together<br />

with those of allozymes and four nuclear genes. Although not <strong>de</strong>tectable with allozymes,<br />

cla<strong>de</strong>-specific allele frequencies of nuclear genes positioned this barrier between 7°25’S<br />

and 14°S. Using a coalescent approach, a secondary contact zone was found with<br />

contrasted migration patterns across the barrier, suggesting that selection against hybrids<br />

acted differentially on each si<strong>de</strong> of the barrier <strong>de</strong>pending on the gene un<strong>de</strong>r scrutiny. Allele<br />

introgression between mt cla<strong>de</strong>s was clearly i<strong>de</strong>ntified and seems to occur in the opposite<br />

direction of migration with the exception of Lysozyme for which a selective sweep is<br />

suspected. In addition, the historical <strong>de</strong>mographic expansion previously proposed in the<br />

South EPR from a multi-species approach was confirmed by the nuclear gene analysis.<br />

176


Chapitre 4<br />

Introduction<br />

Physical barrier to gene flow, and especially tectonically-driven topographic<br />

disruptions (i.e. mountain formation by folding or vertical uplift or continent separation)<br />

are well-known to promote allopatric speciation (e.g. Trewick et al 2000, Cox 2001).<br />

During the process of isolation, mutations are gradually fixed into the divergence, and,<br />

un<strong>de</strong>r the mutation-drift equilibrium, are expected to be randomly distributed over the<br />

whole genome. Contrasted selective pressures on genes, differing evolutionary rates and<br />

the high stochasticity of the coalescence process may however provoke a great<br />

heterogeneity between genes when estimating species divergence across the barrier (e.g.<br />

Barton 1985). Moreover, the strength of a barrier is likely to evolve during the process of<br />

population isolation either by a progressive gene flow attenuation during the formation of<br />

the barrier or by allowing transient secondary contacts across the barrier (see: Avise 2000,<br />

Nielsen & Wakeley 2001). After a period of isolation, this later process often leads to local<br />

hybridization between the previously isolated <strong>de</strong>mes and thus to the subsequent transfer of<br />

alleles between genomes by backcrosses (Barton & Hewitt 1989, Harrison 1990, Rieseberg<br />

et al 1999). Allele introgression however is known to greatly vary between genes<br />

<strong>de</strong>pending on their position from isolation genes and thus the strength of selection against<br />

hybrids (Barton 2000). For a given mutation rate, isolation genes (or genes linked to them)<br />

should thus display lesser introgression than neutral genes. Moreover, sex-biased isolation<br />

mechanisms (such as differing levels of fitness associated with the orientation of the<br />

male/female crossing) can also be responsible for discrepancies of the barrier permeability<br />

between maternally, paternally and double-inherited loci (see Avise 1994). For example,<br />

the mitochondrial isolation of the sperm whale Physeter macrocephalus between oceans,<br />

although not <strong>de</strong>tected in nuclear genes, was explained by female philopatry (Lyrholm et al<br />

1999). Consequently, positioning ‘recent’ barrier to gene flow represents a difficult task,<br />

especially when habitat is ephemeral, highly fragmented or when the barrier can be<br />

affected by tectonic rearrangements: a situation typifying oceanic ridges and its subsequent<br />

vent fauna (e.g. Grassle 1985, Fustec et al 1987, Shank et al 1998, Jollivet et al 1999).<br />

Multiple species comparison of phylogeography patterns represents a powerful tool<br />

to <strong>de</strong>tect shared vicariant events, and to locate more precisely a barrier to gene flow if<br />

dispersal capacities vary slightly across species (e.g. Maggs et al 2008). However, such a<br />

strategy often leads to the <strong>de</strong>tection of a more-or-less wi<strong>de</strong> transition zone associated with<br />

clinal distributions of diagnostic alleles (e.g. Dawson 2001, Thiel et al 2007, Plouviez et al<br />

2009). To such an extent, cross-gene comparisons within a single species may help in<br />

177


Chapitre 4<br />

positioning more precisely the break and to <strong>de</strong>cipher the role of migration versus selection<br />

against hybrids. Polymorphism to divergence analyses on unlinked genes from different<br />

genomes and/or genome compartments (i.e. maternally-inherited mitochondrial versus<br />

nuclear loci) could thus provi<strong>de</strong> useful additional information about the strength of a<br />

barrier, its geographic range and directionality. This was particularly obvious in the case of<br />

the North/South genetic break reported for the <strong>de</strong>ep-sea hydrothermal vent fauna along the<br />

East Pacific Rise (Hurtado et al 2004, Won et al 2003, Plouviez et al 2009). Even if a<br />

transition zone was <strong>de</strong>tected between the northern and southern parts of the East Pacific<br />

Rise (EPR), some species: the Alvinella pompejana polychaete, for example, displays a<br />

clear mitochondrial divergence across the Equator (Hurtado et al 2004, Plouviez et al<br />

2009) whereas others: the bivalve Bathymodiolus thermophilus shows a clinal distribution<br />

of haplotypes associated with a more permeable barrier. This latest species exhibited two<br />

mtCOI divergent cla<strong>de</strong>s overlapping along the North EPR but not in the South (Won et al<br />

2003, Plouviez et al 2009). This result together with the observation that populations from<br />

13°50N also coinci<strong>de</strong>s with the northern range limit of the species, questioned about two<br />

alternative colonization scenarii: (1) the emergence of a secondary contact zone in the<br />

North after the setting up of a North/South barrier, or (2) the migration of only one<br />

mitochondrial lineage to the South through this barrier. The precise location of this<br />

North/South barrier and its role in the population distribution of the <strong>de</strong>ep-sea mussel B.<br />

thermophilus needs thus a more in-<strong>de</strong>pth investigation.<br />

The bivalve Bathymodiolus thermophilus are known to possess great dispersal<br />

capabilities. In<strong>de</strong>ed, millions of oocytes are released per female and fertilized in <strong>de</strong>ep-sea<br />

water over the mussel beds to <strong>de</strong>velop into small 60-µm planktotrophic embryos (Le<br />

Pennec 1984). Recent studies on closely-related species <strong>de</strong>monstrated that these<br />

planktotrophic larvae are able to reach upper-lighted layers of the ocean for feeding to<br />

finally reach a size of about 400 µm prior to metamorphosis and settlement (Mullineaux et<br />

al 2005, Arellano & Young 2009). Because <strong>de</strong>ep currents located several hundreds metres<br />

above the bottom can be twice faster than on the seafloor, and, often very different in<br />

direction from those of the sea surface (Thomson et al 1990, Cannon & Pashinski 1997),<br />

vertical position of larvae is critical for dispersal between sites (Mullineaux et al 2005),<br />

particularly with a planktonic larval duration up to 8 months (as estimated in B. childressis,<br />

Arellano & Young 2009). Such dispersal particularities could thus favour population<br />

connectivity along previously isolated ridges and could explain both the greatest<br />

178


Chapitre 4<br />

permeability of ‘ancient’ topographic barrier and the given directionality of gene flow<br />

along the ridge.<br />

In the present study, a comparison of mtCOI haplotypes, allozymes and alleles at 4<br />

nuclear genes will address several questions about the evolution of mussel populations<br />

located around the North/South barrier of the East Pacific Rise which are:<br />

(1) are allele distributions at other loci consistent with the 7°25’S-14°S barrier to<br />

gene flow proposed using mtCOI (Plouviez et al 2009)?<br />

(2) how great is the permeability of this geographic barrier? and could gene flow be<br />

really asymmetric between the North and South regions?<br />

(3) do the two divergent mtCOI lineages able to cross-hybridized? and finally,<br />

(4) does the hypothesis of population expansion in the south EPR still hold when<br />

using nuclear genes on Bathymodiolus?<br />

To answer these questions, alleles of four nuclear loci were cloned and sequenced<br />

using the mark-recapture (MR) cloning protocol previously <strong>de</strong>veloped by Bierne et al<br />

(2007) using several individuals of 7 disjunct populations located along the whole EPR.<br />

Sequence patterns were then compared with mitochondrial haplotype and allozyme ones<br />

and several evolutionary hypotheses tested using a coalescent approach.<br />

179


Chapitre 4<br />

Material and Methods<br />

Collection<br />

Bathymodiolus thermophilus specimens were sampled from seven <strong>de</strong>ep-sea<br />

hydrothermal vent fields along the East Pacific Rise (Table 4.8) using the tele-manipulated<br />

arm of the manned submersible Nautile operated from the oceanographic vessels Le Nadir<br />

and L’Atalante during two oceanographic cruises: at 9°50’N during HOT 1996 and from<br />

7°25’S to 21°33’S during BIOSPEEDO 2004. During both cruises, all fresh specimens<br />

were measured and dissected on board and tissues (mantle and muscles) were preserved in<br />

80% alcohol. In addition, the anterior muscle of each individual was also frozen in liquid<br />

nitrogen for allozyme analyses during BIOSPEEDO 2004.<br />

Table 4.8 Location and sample size of populations sampled for allozymes (n allo ),<br />

mtCOI (n mtCOI ), and nuclear genes.<br />

n SAHH , n Lyso , n Sulfo1 , n Sulfo2 , total number of recaptured alleles for each nuclear gene. In bold<br />

number of recaptured individuals.<br />

Vent Geographical<br />

field position<br />

9°50N 9°50’N<br />

104°17’W<br />

7°25S 7°25’S<br />

107°47’W<br />

14°00S 13°59’S<br />

112°29’W<br />

17°25’S 17°25’S<br />

113°12’W<br />

17°34’S 17°35-36’S<br />

113°15’W<br />

18°33S 18°33’S<br />

113°24’W<br />

21°33S 21°33’S<br />

114°18’W<br />

Depth<br />

(m)<br />

n allo n mtCOI n SAHH n Lyso n Sufo1 n Sulfo2<br />

2530 0 45 10 16 14 10<br />

6 11 9 8<br />

2735 10 48 33 29 22 12<br />

25 19 17 11<br />

2623 38 30 19 24 15 12<br />

12 14 12 10<br />

2575 137 21 16 15 14 12<br />

11 10 12 8<br />

2595 23 60 15 24 14 14<br />

10 15 12 10<br />

2636 33 30 17 19 15 9<br />

11 12 12 7<br />

2800 103 27 28 24 13 9<br />

19 14 8 6<br />

180


Chapitre 4<br />

Allozyme genotyping<br />

Eight enzyme systems were genotyped on B. thermophilus from 7°S to 21°S (Table<br />

4.8) following the protocols of Pasteur et al (1987). Proteins from each frozen sample of<br />

adductor muscle were extracted according to Piccino et al (2004)’s method. Enzyme<br />

electrophoreses were conducted on 12% starch gel using two different buffer systems: (1)<br />

Tris-Citrate pH 8.0 (TC 8) for Phosphoglucomutase (Pgm, E.C. 5.4.2.2), Mannose<br />

phosphate isomerase (Mpi, E.C. 5.3.1.8), Octopine <strong>de</strong>shydrogenase (Odh, 1.5.1.11) and<br />

Leucine amino peptidase (Lap, 3.4.11.1); (2) Tris-Citrate pH 6.7/6.3 (TC 6.7) for Glucose<br />

phosphate isomerase (Gpi, 5.3.1.9), Malate <strong>de</strong>shydrogenase-1 and -2 (Mdh high and low,<br />

1.1.1.37), and Hexokinase-1 (Hk, 2.7.1.1). Loci were numbered according to the<br />

<strong>de</strong>creasing anodal electromorph mobility in multi-loci systems by calibration with B.<br />

azoricus genotypes from Boutet et al (2009).<br />

DNA sequencing<br />

Genomic DNA was extracted using a CTAB-PVP extraction procedure following<br />

Jolly et al (2006)’s protocol. Mitochondrial lineages of B. thermophilus were i<strong>de</strong>ntified<br />

from Cytochrome Oxidase I gene (mtCOI) sequences previously obtained by Plouviez et al<br />

(2009: Table 4.8). Sequences from 4 nuclear genes were obtained from the same<br />

individuals using the mark-recapture (MR) cloning technique <strong>de</strong>veloped by Bierne et al<br />

(2007) with a capture effort of X2, using primers <strong>de</strong>veloped by Faure et al (2009) for the<br />

intron-exons gene portions of S-A<strong>de</strong>nosyl Homocystein Hydrolase (SAHH) and Lysozyme<br />

(Lyso) and TCTTTAAAGTCAGGATCACATTGG (5’-3’ Forward) and<br />

TAAGGCAAAGTGGAACAACGAGACCGC (5’-3’ Reverse) for the two other exonic<br />

paralogous genes of the Sulfotransferase (called Sulfo1 and Sulfo2, respectively). PCRamplification<br />

were conducted in a 20-µl solution that inclu<strong>de</strong>d 1x buffer (supplied by<br />

manufacturer), 2 mM MgCl 2 , 0.25 mM (0.3 mM for Sulfotransferase) each dNTPs, 0.4 µM<br />

each primers, 0.5 U of Taq polymerase (Thermoprime plus), 0.5µl of 2%-BSA, 2 µl of<br />

template DNA and sterile H 2 O. Thermal cycling parameters used an initial <strong>de</strong>naturation at<br />

94°C for 5 min, followed by 35 cycles at 94°C for 35s, 60°C for 60s and 72°C for 2min,<br />

before a final 10 min extension at 72°C. Each individual was PCR-amplified separately<br />

with a different set of 5’ tailed primers (4-nucleoti<strong>de</strong>s tail).<br />

181


Chapitre 4<br />

Statistical Methods<br />

The program Genetix 4.05.2 (Belkir et al 2004) was used to perform genetic<br />

population analyses on allozymes. For each locus, allele frequencies, heterozygosities and<br />

Weir & Cockerham (1984)’s f estimator (<strong>de</strong>parture from Hardy-Weinberg equilibrium<br />

tested by a 1000-permutations test) were estimated for each population along the EPR. The<br />

overall genetic differentiation across populations was estimated in all loci using the Weir<br />

& Cockerham (1984)’s θ estimator and tested from being significantly different from zero<br />

by 1000 permutations. The isolation-by-distance mo<strong>de</strong>l was then tested with a Mantel Z<br />

test using Genetix 4.05.2.<br />

DNA sequences obtained from the MR method were proofread using CodonCo<strong>de</strong><br />

Aligner 2.0.6 (http://www.codonco<strong>de</strong>.com/aligner/). Sequence alignments were initially<br />

performed with ClustalW (Thompson et al 1994) using BioEdit version 6.0.6 (Hall, 1999)<br />

and improved manually. Number of recaptured alleles varied from one population to<br />

another and allele sample size of each gene and population is indicated in Table 4.8. For<br />

nuclear genes, in vitro recombinants between different individuals from the same PCR set<br />

were easily <strong>de</strong>tected thanks to abnormal combination of the 5’-tails and removed. Multiple<br />

recaptures also allowed us to discard intra-individual in vitro recombinants and putative<br />

artefactual/somatic mutations.<br />

Because of the random nature of the recapture of cloned sequences, it was not<br />

possible to distinguish ‘true’ homozygotes from heterozygotes with a few recaptured<br />

alleles. Consequently, one recaptured allele was randomly sampled from each individual to<br />

avoid sample bias when performing <strong>de</strong>mographic analyses and genetic diversity<br />

estimations (see: Table 4.8). For each locus, genealogical relationships among alleles were<br />

estimated using the median joining algorithm of the Network software (version 4.5.0.0;<br />

www.fluxus-engineering.com: Ban<strong>de</strong>lt et al 1999) to <strong>de</strong>tect potential divergent cla<strong>de</strong>s.<br />

Geographic distributions of divergent cla<strong>de</strong>s were then examined to locate potential barrier<br />

to gene flow, by plotting synthetic cla<strong>de</strong>-specific allele-frequency distributions for each<br />

locality. These cla<strong>de</strong>s were thus <strong>de</strong>fined as mainly North (MN) or South (MS) according to<br />

these allele-frequency distributions. Localities from each si<strong>de</strong> of the putative barrier were<br />

grouped together to test for genetic differentiation across this barrier using F-statistics (φ st ;<br />

182


Chapitre 4<br />

Hudson et al 1992) computed via DNAsp 4.10.3 (Rozas et al 2003) and tested by a 1000-<br />

permutations test.<br />

With the same software, nucleoti<strong>de</strong> diversity (π n ) and Watterson’s theta (θ w ) were<br />

then estimated within both MN and MS cla<strong>de</strong>s. A <strong>de</strong>mographic event, should impact the<br />

whole genome in the same way. Conversely, selective effects could act only on few genes.<br />

To disentangle the effects of selective vs <strong>de</strong>mographic effects, <strong>de</strong>parture to neutral<br />

evolution (i.e. no selection and no <strong>de</strong>mographic event) was thus tested for each gene both<br />

in MN and MS cla<strong>de</strong>s by Tajima’s (D, 1989) and Fu & Li’s (F, 1993) statistics.<br />

Because coalescent mo<strong>de</strong>l assumes no recombination, absence of recombination<br />

was thus assessed before proceeding all following analyses using the four-gamete test<br />

(Hudson & Kaplan 1985). A multi-locus Hudson-Kreitman-Aguadé test (HKA, 1987) was<br />

performed separately for MN and MS B. thermophilus cla<strong>de</strong>s with one sequence of B.<br />

azoricus as outgroup via the software HKA (J. Hey’s web page:<br />

http://lifesci.rutgers.edu/∼heylab/HeylabSoftware.htm#HKA). This test compares<br />

divergence and polymorphism at several loci to <strong>de</strong>tect if some of these loci display<br />

<strong>de</strong>parture to neutral evolution. Discrimination between bottleneck versus selective sweep<br />

was also tested separately in MN and MS cla<strong>de</strong>s using all genes, except Sulfo2 which did<br />

not show any divergent cla<strong>de</strong> and was thus separated according to the geography, using<br />

sweep_bott (Galtier et al 2000) by 100 000 first step iterations, 1 000 000 second iterations<br />

and 20 optimizations processes with a theta range from 1 to 30. Because this software<br />

assumed Infinite Site mutation mo<strong>de</strong>l and no reverse substitution, sites that did not follow<br />

this assumption were discar<strong>de</strong>d from the dataset. Mc Donald-Kreitman (1991) tests were<br />

not found pertinent because of the absence of fixed non-synonymous mutations between B.<br />

thermophilus and B. azoricus for all the studied genes.<br />

Migration across the North/South barrier<br />

Migration routes across the barrier were examined via Migrate 3.0.3 (Beerli &<br />

Felsenstein 2001) by looking this time at two geographically differentiated groups of sites<br />

(alleles here separated according to the northern and southern parts of the barrier, whatever<br />

their cla<strong>de</strong> adherence). Unfortunately, because of the Mark-Recapture method, few<br />

individuals (1-6 individuals per site, 3.4 in average) displayed alleles in all genes.<br />

Migration rates and thetas were thus estimated by MCMC method on each gene,<br />

separately. Initial Watterson’s theta (θ w ) was calculated for each group of populations<br />

183


Chapitre 4<br />

using DNAsp 4.10.3 and used as a starting parameter in Migrate 3.0.3 to estimate the gene<br />

flow parameters (θ*M) with MCMC runs of ten short (500 steps, 10 000 sampled<br />

genealogies), three long chains (5 000 steps, 100 000 sampled genealogies) and swapping<br />

procedure among 4 heated chains.<br />

Isolation with migration between the two mt lineages<br />

The 0.9%-divergent geographic mt lineages, previously <strong>de</strong>tected between the<br />

northern and southern parts of the EPR (Won et al 2003, Plouviez et al 2009), were<br />

hypothesized to have diverged about 1.3 million years ago with the formation of the<br />

Gofar/Discovery transform faulting complex (Plouviez et al 2009). Because permeability<br />

of the barrier to gene flow can vary from gene to gene, levels of allele introgression may<br />

differ wi<strong>de</strong>ly across nuclear genes and individuals from both mtCOI lineages. Rates of<br />

migration between the two mtCOI lineages (here migration corresponds in migration<br />

between genome contrary to migration in Migrate-n that is consi<strong>de</strong>red across a physical<br />

barrier) and time since population splitting were thus estimated to <strong>de</strong>tect impact of the<br />

barrier on each nuclear gene by fitting an isolation with migration mo<strong>de</strong>l (IMa and IM<br />

programs, Hey & Nielsen 2004). For each gene, individuals were thus classified into two<br />

groups according to their mitochondrial origin (mtNorth and mtSouth with individuals<br />

having the MN and MS mtCOI lineages, respectively). The HKY mutation mo<strong>de</strong>l<br />

(Hasegawa et al 1985) was chosen because of <strong>de</strong>tection of sites with more than two<br />

character states in the dataset. The method <strong>de</strong>rives analytically posterior probability<br />

distributions from sampled gene genealogies for effective sizes of both the ancestral and<br />

sister cla<strong>de</strong>s (θa θ1 θ2 respectively), the directional gene-flow rates between the <strong>de</strong>rived<br />

lineages (migration according to the IM signification: m1 and m2) and the time elapsed<br />

since the lineages splitting (t). A gene-by-gene estimation of parameters were also<br />

performed using the IM software to search (1) preferred migration orientation of each gene<br />

and (2) whether migration events are recent (introgressed alleles due to a recent secondary<br />

contact) or old (parapatric isolation process). To check convergence of the Markov Chain<br />

Monte Carlo (MCMC) toward the true stationary distribution, multiple runs were<br />

performed using different starting points and autocorrelation between parameter values<br />

was assessed over the course of the runs. A swapping procedure was also used to enhance<br />

the mixing of chains using a geometric heating scheme with 6 Metropolis coupling chains.<br />

Inheritance scalar was assigned to 1 because of the autosomal nature of nuclear genes. To<br />

184


Chapitre 4<br />

convert IMa parameters of divergence time into years, effective population size and<br />

number of introgressed individuals, the geometric mean of mutation rate per year over loci<br />

(U) was estimated using the previous calibration of the mutation rate obtained by Faure et<br />

al (2009), which are also consistent with the calibration across the Isthmus of Panama<br />

when comparing B. thermophilus and B. azoricus. These parameters were converted into<br />

number of individuals within lineage (N), time in years and number of individuals<br />

displaying gene-flow between lineages (M) according to the relationships: N = θ /(4UG),<br />

M = 2Nm, t (years) = t/U, with θ (<strong>de</strong>mographic parameter), m (migration) and t (time)<br />

parameters estimated by IMa and G the number of years per generation.<br />

Hybrid <strong>de</strong>tection using SAHH RFLP analysis<br />

Because the SAHH nuclear gene also displays a strong divergence between two<br />

geographic cla<strong>de</strong>s and appears to be genetically linked with mt lineages, occurrence of<br />

putative introgressed/hybrids between mitochondrial lineages were checked by looking for<br />

individuals that possessed one allele typical from each cla<strong>de</strong> (called N/S individuals). A<br />

RFLP approach was thus done on a fixed site previously found in the divergence between<br />

the two sets of SAHH alleles to <strong>de</strong>tect N/S individuals by screening all individuals from<br />

the Bathymodiolus collection onto a 2% agarose gel. The Hinf I restriction enzyme was<br />

thus applied on the SAHH PCR-product fragments to discriminate individuals with and<br />

without this restriction site. Incubation of PCR-products was done at 37°C during 1h30’<br />

within a 20µl total volume containing 17µl of PCR product, 1X buffer (supplied by the<br />

manufacturer) and 10U of Hinf I (Ozyme TM ).<br />

Testing the hypothesis of diversifying selection between mt lineages<br />

Because of the lack of non-synonymous mutations in the divergence of the two mt<br />

cla<strong>de</strong>s of B. thermophilus or between these cla<strong>de</strong>s and the Atlantic outgroup species B.<br />

azoricus, it was not possible to use the Mc Donald-Kreitman test to search for positive<br />

selection in the separation of the B. thermophilus lineages. We therefore proposed to<br />

employ a new test based on the posterior distribution of the difference between thetas of<br />

the two mt lineages un<strong>de</strong>r the assumption of a structured coalescent into two sister <strong>de</strong>mes.<br />

This tests disruptive selection between two groups relying on the fact that selection acting<br />

185


Chapitre 4<br />

on one si<strong>de</strong> of the barrier at a given gene would lead to different nucleotidic diversities<br />

across the barrier. Observed difference between diversities at a given locus is estimated by<br />

Δθ w (θ w north - θ w south), which un<strong>de</strong>r the mo<strong>de</strong>l of neutral isolation of the <strong>de</strong>mes must be<br />

normally distributed. Genes showing Δθ w outliers are therefore a good candidate for<br />

disruptive selection across the barrier. This test is in<strong>de</strong>pen<strong>de</strong>nt of the mutation rate as<br />

genes should accumulate mutations at the same rate in polymorphism of the two isolated<br />

cla<strong>de</strong>s. Exception occurs if selection acted more efficiently in one part of the barrier and<br />

purged allele diversity, or if populations from a given cla<strong>de</strong> have endured a bottleneck. In<br />

the first case, one can expect that only few outliers must be <strong>de</strong>tected whereas in the second<br />

case all genes must display an outlier value. To simulate Δθ w from a neutral mo<strong>de</strong>l, we<br />

used the HABC software MsBayes recently <strong>de</strong>veloped from Hickerson et al (2006) to<br />

calculate a series of 1000 pairs of thetas for each gene un<strong>de</strong>r the coalescent mo<strong>de</strong>l of<br />

species split from an ancestral population without bottleneck.<br />

186


Chapitre 4<br />

Results<br />

Geographic distribution of allozymes<br />

A series of differentiation tests were performed on B. thermophilus populations<br />

located along the SEPR from 7°25S to 21°33S. Among the eight allozyme loci, the most<br />

common allele displayed a frequency up to 95% at four loci: Mpi, Odh, Mdh low and Hk.<br />

These four loci were thus consi<strong>de</strong>red as monomorphic whereas the remaining 4 loci (Pgm,<br />

Lap, Gpi and Mdh high) were found polymorphic enough to investigate their allele<br />

distribution over the range of the EPR (see supplementary material S1: allele frequencies).<br />

The global test of genetic differentiation did not show significant heterogeneity in allele<br />

frequencies between populations (Fst = 0.007, p value > 0.05), and thus did not allow us to<br />

<strong>de</strong>tect the presence of any genetic structure along the south EPR. Plot of genetic distances<br />

(θ(1-θ)) against geographic distance (Fig. 4.16) and Mantel test (p value > 0.17) also failed<br />

to reveal a trend toward isolation-by-distance.<br />

Fig. 4.16 Distribution of allozyme differentiation (measured by using the distance<br />

theta(1-theta)) against the geographic distance between sites.<br />

187


Chapitre 4<br />

Polymorphism and divergence from DNA sequences<br />

Networks of alleles revealed a pronounced geographic structure of the mussel<br />

Bathymodiolus thermophilus between Northern and Southern EPR. All genes including the<br />

mitochondrial COI were shaped into two geographic cla<strong>de</strong>s separated by a more or less<br />

pronounced divergence (see Fig. 4.17). The SAHH gene revealed the presence of two 2%-<br />

divergent cla<strong>de</strong>s well established from each part of the previously suggested 7°25’S-14°S<br />

mitochondrial break (Plouviez et al 2009, Fig. 4.17). SAHH cla<strong>de</strong> 1 contains most of<br />

sequences from individuals sampled in the northern EPR and 7°25S whereas cla<strong>de</strong> 2<br />

corresponds to individuals only sampled from the other southern EPR sites. Two 0.3%-<br />

divergent cla<strong>de</strong>s were also found in Lysozyme gene but are more difficult to attribute to a<br />

particular region.<br />

188


Chapitre 4<br />

Fig. 4.17 Median Joining Networks on nuclear genes.<br />

SAHH, S-A<strong>de</strong>nosyl Homocystein Hydrolase; Lyso, Lysozyme; Sulfo 1 and 2,<br />

Sulfotransferase paralogues 1 and 2 together with the Cytochrome Oxidase I (mtCOI:<br />

adapted from Plouviez et al 2009).<br />

189


Chapitre 4<br />

However, a 1 bp in<strong>de</strong>l was only found in sequences sampled in the southern EPR<br />

sites below the latitu<strong>de</strong> of 7°25’S (Fig. 4.18), indicating that the 7°25’S-14°S barrier also<br />

played a role in impeding the spread of new mutations. The paralogous gene Sulfo 1 also<br />

displayed a clear geographic structure between the north and the south of the barrier with a<br />

0.5% divergence between the two cla<strong>de</strong>s whereas the less-sampled paralogous Sulfo 2<br />

gene did not reveal any clear structure (Fig. 4.17). This geographic break leads to genetic<br />

differentiation between the two geographic groups of populations (i.e. below and above the<br />

7°25’S-14°S limit) with φ st values significantly different from zero for all sampled genes<br />

(Table 4.9).<br />

SAHH Lyso Sulfo 1<br />

9°N<br />

7°S<br />

14°S<br />

17°25’S<br />

17°34’S<br />

18°S<br />

21°S<br />

120°W 100°W 80°W<br />

1000 km<br />

20°N<br />

10°N<br />

0°<br />

10°S<br />

20°S<br />

30°S<br />

Allele from cla<strong>de</strong>1 (with 1bp ‘insertion’ for Lyso in the intron)<br />

Allele from cla<strong>de</strong>1 (with 1bp ‘<strong>de</strong>letion’ in the intron)<br />

Allele from cla<strong>de</strong>2 (with 1bp ‘insertion’ for Lyso in the intron)<br />

Allele from cla<strong>de</strong>2 (with 1bp ‘<strong>de</strong>letion’ in the intron)<br />

Fig. 4.18 Distribution of divergent alleles along the East Pacific Rise for three<br />

nuclear genes.<br />

SAHH, S-A<strong>de</strong>nosyl Homocystein Hydrolase; Lyso, Lysozyme; Sulfo 1,<br />

Sulfotransferase paralogues 1<br />

190


Chapitre 4<br />

Table 4.9 Summary statistics of nucleoti<strong>de</strong> polymorphism for mitochondrial and<br />

nuclear genes.<br />

N, number of sequences; S number of alleles; θ W , Waterson’s theta; π n , nucleoti<strong>de</strong><br />

diversity; D, Tajima’s in<strong>de</strong>x; F*, Fu & Li in<strong>de</strong>x. These parameters were calculated in MN<br />

and MS cla<strong>de</strong>s (mainly found in the North or in the South of the barrier positioned between<br />

7°25’S-14°S) providing that a fixed divergence was found in the consi<strong>de</strong>red gene. Because<br />

Sulfo2 did not display divergent cla<strong>de</strong>s, these parameters were calculated in each part of<br />

the barrier, respectively.<br />

φ st values correspond to levels of differentiation between the Northern and Southern<br />

populations located from each part of the barrier.<br />

*: P< 0.05, **: p


Chapitre 4<br />

Tajima’s D and Fu & Li’s F* in<strong>de</strong>xes estimated from MN and MS cla<strong>de</strong>s showed<br />

significantly negative values for all genes in the south of the barrier (Table 4.9) whereas<br />

<strong>de</strong>partures were only <strong>de</strong>tected at three loci (Lyso and Sulfo 1 and 2) in the MN cla<strong>de</strong>.<br />

Moreover, levels of significance of the Fu & Li’s tests, which are more sensitive to<br />

<strong>de</strong>mographic events, were similar or lesser than for Tajima’s tests when performed for the<br />

North cla<strong>de</strong>. The multi-locus HKA test performed using all genes except Sulfo2, showed a<br />

significant <strong>de</strong>parture to neutral evolution (P = 0.0004) in both the MN and MS cla<strong>de</strong>s,<br />

indicating that at least one gene could be un<strong>de</strong>r selection. Sulfo 1 displayed the highest<br />

divergence and polymorphism <strong>de</strong>viation when compared to other genes. When Sulfo 1 was<br />

<strong>de</strong>leted from the dataset, the HKA test conformed to the mo<strong>de</strong>l of neutral evolution (P =<br />

0.2812). This gene displayed a polymorphism/divergence ratio of 8.66: a value that is more<br />

than eight times greater than the expected one (1.05) with a very low divergence between<br />

B. thermophilus and B. azoricus, a characteristic of balancing or purifying selection. In the<br />

Northern cla<strong>de</strong>, likelihood ratio tests (LRT) performed by the Sweep_bott analysis (Table<br />

4.10) also revealed <strong>de</strong>partures to the neutral hypothesis (neutral mo<strong>de</strong>l M1 was<br />

significantly different from the bottleneck M2 and selective sweep mo<strong>de</strong>l M3) but did not<br />

allow to choose between the two last hypotheses. However, when looking at genes<br />

separately, the Lysozyme was the only gene displaying a significant LRT between M1 and<br />

M3 both in Northern and Southern cla<strong>de</strong>s. This indicated that M3 should be more probable<br />

than M2 with a mo<strong>de</strong>rate selective sweep on Lyso. In the South, LRT also showed that M2<br />

and M3 should be more appropriate than M1 (Table 4.10). Similarly, M3 better explained<br />

sequence datasets when compared to M2, and indicated the occurrence of, at least, two<br />

<strong>de</strong>partures to neutral expectation (sweep and/or bottleneck) in the South. When looking at<br />

genes more specifically, the only gene that conformed to the neutral coalescent was Sulfo<br />

2. The two remaining loci displayed significant LRTs whereas Sulfo1 was not used<br />

because of its lack of structure (only one highly-frequent allele surroun<strong>de</strong>d by <strong>de</strong>rived<br />

singletons). Moreover, the very low sample size of Sulfo 2 questioned the significance of<br />

the test (lack of power) in retaining alternative mo<strong>de</strong>ls suggesting that the bottleneck (M2)<br />

mo<strong>de</strong>l could be valid in the southern cla<strong>de</strong>.<br />

192


Chapitre 4<br />

Table 4.10 Likelihoods and likelihood ratio test (LRT) computed via sweep_bott to<br />

discriminate selective sweep to bottleneck in Northern and Southern EPR.<br />

M1, neutral mo<strong>de</strong>l; M2, bottleneck mo<strong>de</strong>l; M3, selective sweep mo<strong>de</strong>l; NS, nonsignificant.<br />

Significances of the LRT were tested according to a χ2 table.<br />

Likelihood SAHH Lyso Sulfo1 Sulfo2<br />

North<br />

M1 -138.255 -40.107 -52.684 -27.066 -18.398<br />

M2 -132.459<br />

M3 -129.096 -39.301 -48.558 -25.554 -15.683<br />

LRT<br />

(significance)<br />

1.612 (NS) 8.252 (2.5%) 3.024 (NS) 5.430 (NS)<br />

M2-M1: 11.518 (0.5%)<br />

M3-M1: 18.312 (2.5%)<br />

M3-M2: 6.726 (NS)<br />

South<br />

M1 -135.190 -56.247 -60.755 - -18.187<br />

M2 -117.380<br />

M3 -103.950 -46.785 -40.955 - -16.210<br />

LRT<br />

(significance)<br />

M2-M1: 35.619 (0.5%)<br />

M3-M1: 62.480 (0.5%)<br />

M3-M2: 26.861 (0.5%)<br />

18.924 (0.5%) 39.600 (0.5%) - 3.954 (NS)<br />

Migration across the geographical barrier<br />

To test for the permeability of the barrier itself, migration rates between the<br />

northern and southern parts of the EPR after positioning it between 7°25’S and 14°S were<br />

estimated using the software Migrate_n. Because very few individuals got a multiloci<br />

genotype, runs were performed for each nuclear gene, separately. As previously shown<br />

from Migrate_n computations on the mtCOI (Plouviez et al 2009), migration across the<br />

barrier was clearly asymmetrical with direction <strong>de</strong>pending on the gene un<strong>de</strong>r scrutiny<br />

(Table 4.11). The genes SAHH and Sulfo1 however displayed migration rates very close to<br />

zero, suggesting that these two genes are exchanging very few alleles across the barrier.<br />

Several introgressed SAHH alleles coming from the North cla<strong>de</strong> were however recaptured<br />

in individuals of the south cla<strong>de</strong> at 14°S, suggesting that a low migration occurs from<br />

North to South for this gene.<br />

193


Chapitre 4<br />

Table 4.11 Migration rates across the 7°25’S-14°S barrier using the software<br />

Migrate_n.<br />

N/S, North to South; S/N, South to North; θ N , Theta in the North group; θ S , Theta in the<br />

South group; MLE, Maximum-likelihood Estimates; P, percentile.<br />

Loci 4Nm N/S 4Nm S/N θ N θ S<br />

SAHH<br />

MLE 1.28 1.59 10 -12 0.007 0.032<br />

P 0.005 0.18 7.96 10 -13 0.003 0.017<br />

P 0.995 4.03 0.49 0.015 0.054<br />

Lyso<br />

MLE 36.56 0.23 0.008 0.026<br />

P 0.005 22.63 3 10 -3 0.006 0.016<br />

P 0.995 54.63 1.43 0.012 0.048<br />

Sulfo1<br />

MLE 4.74 10 -11 1.39 0.028 0.021<br />

P 0.005 2.37 10 -11 0.14 0.015 0.014<br />

P 0.995 0.83 4.76 0.052 0.033<br />

Sulfo2<br />

MLE 4.59 10 -8 20.85 0.048 0.030<br />

P 0.005 2.30 10 -8 3.29 0.006 0.014<br />

P 0.995 2 10 -4 41.11 242.50 0.044<br />

Detection of N/S individuals using the SAHH gene<br />

Because of its strong divergence across the 7°S-14°S barrier, the nuclear gene SAHH was<br />

choosen to quantify the level of N/S individuals within populations over the whole EPR. A<br />

RFLP analysis on a fixed substitution in the divergence of the SAHH gene between the<br />

geographic regions revealed that only 13 individuals displayed at least one nuclear allele<br />

corresponding to the North cla<strong>de</strong> in the South EPR. Except one individual showing two<br />

Northern type alleles in the South, all the remaining individuals possessed both the<br />

Northern and Southern type alleles. N/S individuals were mainly located at 14°S and<br />

<strong>de</strong>creased abruptly in frequency further south (Fig 4.19). None were <strong>de</strong>tected at 7°25S and<br />

further North. When looking at the southern mt lineage, N/S individuals were not<br />

associated with a specific haplotype: the Northern type alleles in SAHH being found both<br />

in MN and MS mtCOI lineages whereas the Southern one was only restricted to the MS<br />

mtCOI lineage.<br />

194


Chapitre 4<br />

Fig 4.19 Detection of individuals with both North and South type alleles in SAHH or<br />

allele type that did not correspond in geographic position across the barrier using<br />

RFLP analyses.<br />

Testing the hypothesis of Isolation with migration between mt lineages<br />

In contrast to the Migrate_n analysis performed on groups of populations (testing<br />

the strength of the physical barrier), the isolation with migration mo<strong>de</strong>l was tested between<br />

the two main mitochondrial lineages (MN, MS), these lineages being mixed over the whole<br />

southern EPR. This analysis was done to estimate the level of allele introgression between<br />

the two previously isolated cla<strong>de</strong>s. Mo<strong>de</strong>l parameters estimated by IMa (θa, θ1, θ2, t, m1<br />

and m2) are presented in Table 4.12.<br />

195


Chapitre 4<br />

196


Chapitre 4<br />

Divergence time between the two mt lineages displayed the highest likelihood<br />

values around 1.3 Mya (Fig. 4.20). The southern lineage has a greater population size than<br />

those of the ancestral lineage and the northern one (Fig. 4.20), suggesting a population<br />

expansion of the MS lineage. The multilocus estimates of m1 (=0.055) and m2 (=0.025)<br />

parameters showed an asymmetric migration from MN to MS lineages (Table 4.12, Fig.<br />

4.20). Because population size of the MN lineage was lower than that of the MS lineage,<br />

effective migration rate per generation from MN to MS was lower than in the other sense<br />

(Table 4.12).<br />

197


Chapitre 4<br />

A<br />

B<br />

C<br />

Fig 4.20 Marginal posterior probability distribution for divergence time between<br />

lineages (A) effective population sizes of the ancestral population and its two <strong>de</strong>rived<br />

ones (B) and directional gene-flow rates between the two mt lineages (C), estimated<br />

by the isolation with migration mo<strong>de</strong>l IMa.<br />

Parameters correspond in those from the Table 4.12.<br />

198


Chapitre 4<br />

The gene-to-gene approach confirmed this asymmetric migration sense. In<strong>de</strong>ed,<br />

three nuclear genes (Lyso, Sulfo1, Sulfo2) displayed non-null values of m1 whereas m2<br />

peak was found to be zero (Table 4.13, Fig 4.21). SAHH gene showed a m1 value close to<br />

zero (however, the non logarithmic curve shape could indicate few hybrids) but a m2 value<br />

slightly greater than m1. Values of the mean time of gene flow events were estimated<br />

between 2000 and 391000 years, and were thus much more recent than the 1.3 My value<br />

estimated as a starting point for the isolation of the mtCOI lineages (Table 4.13, Fig 4.21).<br />

Table 4.13 Levels of migration across lineages (m) and maximum of the mean time of<br />

migration events (mtime) obtained by IM analyses out of 5 loci.<br />

(1) and (2) correspond to MN and MS, respectively. m1, migration from MN to MS and<br />

m2: migration from MS to MN.<br />

Locus m1 m2 mtime1 (million) mtime2 (million)<br />

COI 0 0 - -<br />

SAHH 0 1 - 0.011<br />

Lyso 2 0 0.391 -<br />

Sulfo1 1 0 0.002 -<br />

Sulfo2 1 0 0.020 -<br />

199


Chapitre 4<br />

A<br />

B<br />

Fig. 4.21 IM single locus probability distributions of hybridization rate (m) for each<br />

5 loci for North to South (A) and South to North (B).<br />

200


Chapitre 4<br />

Testing the hypothesis of diversifying selection across the barrier<br />

Estimations of Δθ between Northern and Southern cla<strong>de</strong>s were distributed within<br />

the 95% confi<strong>de</strong>nce envelope obtained for all tested genes when simulating a structured<br />

coalescent following a classical mo<strong>de</strong>l of allopatric isolation (Fig. S2, Supplementary<br />

material). Observed differences between cla<strong>de</strong> diversities were therefore not significantly<br />

different from those expected by the isolation mo<strong>de</strong>l without bottleneck and therefore do<br />

not support assumptions about either diversifying selection acting between MN and MS<br />

cla<strong>de</strong>s or about recent <strong>de</strong>mographic changes on one si<strong>de</strong> of the barrier. mtCOI was the only<br />

gene yielding a Δθ value in a probability class closest to the 5% threshold.<br />

201


Chapitre 4<br />

Discussion<br />

Geographic position of a barrier to gene flow<br />

One-dimensional but highly fragmented habitats such as <strong>de</strong>ep-sea hydrothermal<br />

vents represent an i<strong>de</strong>al mo<strong>de</strong>l to test for stepping-stone or isolation-by-distance mo<strong>de</strong>ls of<br />

populations (Vrijenhoek 1997). However, hydrothermal activity is also ephemeral and<br />

moves along the ridge axis <strong>de</strong>pending on the dynamics of the un<strong>de</strong>rneath magmatic<br />

chamber and its <strong>de</strong>rived heat convection (Watremez & Kervévan 1990, Kervévan 1991).<br />

This leads to massive fauna extinction at the scale of a given vent field (Haymon et al<br />

1991, Tunnicliffe et al 1997, Marcus et al 2009) able to seriously alter expectations of such<br />

population mo<strong>de</strong>ls (Jollivet et al 1999). Because of its unpredictable nature and the level of<br />

fragmentation, one would expect that vent species should <strong>de</strong>velop high dispersal<br />

capabilities. However, most vent species exhibit internal fertilization of eggs and a<br />

lecithotrophic <strong>de</strong>velopment of the larva typifying short-term dispersal ability (Chevaldonné<br />

et al 1997, Tyler and Young, 1999, Jollivet et al 2000, Marsh et al 2001). Paradoxically,<br />

genetic structures that fit the isolation-by-distance mo<strong>de</strong>l were very scarce and often<br />

related to the presence of at least one geographic barrier (cf the Hess Deep triple junction:<br />

see France et al 1992, Black et al 1994, Jollivet et al 1995, Vrijenhoek 1997). On the<br />

contrary, many species, such as the polychaete Alvinella pompejana (Jollivet et al 1995),<br />

the bivalve Bathymodiolus thermophilus (Craddock et al 1995) or the gastropods<br />

Lepetodrilus pustulosus and Eulepetopsis vitrea (Craddock et al 1997) displayed a lack of<br />

geographic structure and were supposed to follow an island-like mo<strong>de</strong>l of dispersal<br />

suggesting great dispersal capabilities.<br />

Species from the <strong>de</strong>ep-sea Bathymodiolus genus has thus been previously <strong>de</strong>scribed<br />

as a challenger for long-distance dispersal. For example, occurrence of amphi-Atlantic<br />

species with virtually no genetic divergence, suggested that present gene flow may still<br />

occur across the Atlantic (Olu-LeRoy et al 2007). Absence of genetic differentiation was<br />

also reported at the scale of the Gulf of Mexico between populations from the Mississippi<br />

Canyon and Alaminos Canyon, 550 km apart, by Carney et al (2006) for the seep species<br />

B. childressi. On the EPR, Craddock et al (1995) also showed a lack of isolation-bydistance<br />

for B. thermophilus when comparing allozyme distributions between the Northern<br />

EPR and the Galapagos sites, and subsequently conclu<strong>de</strong>d about exten<strong>de</strong>d dispersal<br />

capabilities of the mussel.<br />

202


Chapitre 4<br />

However, this latter conclusion was recently challenged by Won et al (2003) and<br />

Plouviez et al (2009), who reported the presence of a genetic break between the Southern<br />

and the Northern parts of the EPR. These latter authors in<strong>de</strong>ed suggested that the abrupt<br />

change in the frequencies of mitochondrial haplotypes was mainly due to a barrier to gene<br />

flow that should be attributable to either strong westward cross-axis currents (Won et al<br />

2003) or the formation of the Discovery/Gofar transform fault complex (Plouviez et al<br />

2009). Precise location of such a barrier was difficult to estimate because of the presence<br />

of a clinal distribution of the three main mtCOI haplotypes (Plouviez et al 2009) but it was<br />

suggested occurring between 7°25S and 14°S. In the present study, allozyme loci failed to<br />

display significant genetic differentiation between 7°25S and the most southern sites, and<br />

thus were in agreement with previous results on the North EPR. On the contrary, the<br />

distribution of two sets of divergent alleles at four nuclear genes confirmed the 7°25’S-<br />

14°S position of a barrier with significant differentiation (φ st ) values in all genes between<br />

these two regions. Moreover, this geographic break also corresponds to a strong<br />

divergence between two cla<strong>de</strong>s of alleles in the SAHH gene. A similar north/south<br />

distribution of cla<strong>de</strong>s starting from 7°25S is also <strong>de</strong>picted in the gene Sulfo1 but with a<br />

very low divergence between these two groups of alleles. Two divergent cla<strong>de</strong>s are also<br />

observed in the gene Lyso but these two sets of alleles are well mixed within populations<br />

over the whole EPR. A 1bp-in<strong>de</strong>l was however observed only restricted to the south of the<br />

barrier for both allelic cla<strong>de</strong>s indicating that alleles from the two cla<strong>de</strong>s has recombined in<br />

the South and that the barrier is nearly impermeable for newly appeared mutations, at least<br />

for this gene.<br />

The isolation with migration mo<strong>de</strong>l revealed a 1.36 My divergence time between the MN<br />

and MS lineages. These results are consistent with the dating of the ∼1.3 Mya splitting<br />

event occurring between Northern and Southern EPR estimated from the multi-species<br />

approach (Plouviez et al 2009), and hypothesised to be due to the formation of a multiple<br />

transform fault complex about 1-2 Mya at these latitu<strong>de</strong>s (Kureth & Rea 1981, Naar & Hey<br />

1989, Francheteau et al 1990). However, although the geographic location of the barrier<br />

was difficult to position using mitochondrial markers (Plouviez et al 2009), the use of<br />

multiple nuclear loci allowed here to locate this barrier more precisely between 7°25’S and<br />

14°S. A strong westward flow of He-3 centered at 15°S (Lupton & Craig 1981, Lupton<br />

1998) was hypothesized to mimic a strong anticyclonic circulation across the Southern<br />

EPR (Fujio & Imasato 1991). If this <strong>de</strong>ep water-masse circulation could have primarily<br />

reinforced isolation between Northern and Southern EPR, putative fluctuations in the<br />

203


Chapitre 4<br />

strength and orientation of the water movements could have also played a crucial role in<br />

allowing transient genetic exchanges across the barrier, particularly for species capable of<br />

reaching the upper water column, such as Bathymodiolus (Arellano & Young 2009).<br />

Evolutionary forces at play around the barrier, however pose questions about the evolution<br />

of the allopatric separation of the two lineages. In<strong>de</strong>ed, if gene flow is impe<strong>de</strong>d by<br />

transform faults, then one should expect that the isolation process will be progressively<br />

reinforced by the axis off-set. Most of the present allele distributions would therefore be<br />

due to past gene flow: a situation that fits well with IM mean time of gene flow events for<br />

the Lyso gene and the stop of the 1bp in<strong>de</strong>l mutation spread to the northern EPR.<br />

Conversely, if gene flow is impe<strong>de</strong>d by strong transversal currents, then sporadic<br />

exchanges are expected and allow a transient secondary contact zone between 7°25S and<br />

14°S: a situation well <strong>de</strong>picted by the mt haplotypes admixture in the North and the<br />

presence of SAHH N/S individuals at 14°S and 17°25S and Sulfo 1 and 2. However,<br />

migration between genomes seems to have occurred in two very different periods of time<br />

since the splitting of the two mt lineages. In<strong>de</strong>ed, mean time of migration events (mtime)<br />

obtained by IM were much more ancient for the gene Lyso (about 400 000 years) than for<br />

the others genes for which mtime are close to present.<br />

Permeability of the barrier and subsequent introgression<br />

Permeability of the equatorial barrier was already proposed by Plouviez et al (2009)<br />

because of the lack of reciprocal monophilies between northern and southern lineages of<br />

several taxa when using mtCOI. In Bathymodiolus, the two divergent mtCOI lineages were<br />

mixed in populations of the northern EPR and displayed, at least, a clinal distribution for<br />

the southern lineage (about 0.25 in the northern populations), suggesting a secondary<br />

contact in the North following the geographic isolation. Such a haplotype admixture was<br />

not <strong>de</strong>tected in the Atlantic <strong>de</strong>spite extensive allele introgressions between nuclear<br />

genomes of both B. puteoserpentis and B. azoricus species (Faure et al 2009). In the EPR,<br />

the geographic location of the secondary contact zone seems however to vary among<br />

studied genes with an ancient gene flow but a lack of ongoing gene flow in the Lyso gene,<br />

whereas a weak secondary contact occurred in the South for the SAHH gene (hybrids only<br />

found from 14°S to 17°34’S) and in the North for the mtCOI gene (two lineages in the<br />

North). These findings may therefore suggest that larval dispersal must be mostly<br />

orientated from South to North (orientated water masse circulation) but selection against<br />

204


Chapitre 4<br />

hybrids may be also more severe in the northern EPR. Differences in the strength of<br />

selection against hybrids from each part of the barrier can thus explain discrepancies<br />

between loci when dispersal is asymmetric (e.g. Barton & Hewitt 1985). Present results<br />

using Migrate_n computations confirmed such a trend with 2 locus favouring migration<br />

from South to North (a situation similar to that of the mitochondrial marker) and 2 locus<br />

favouring migration in the opposite direction. Interestingly loci favouring the South to<br />

North direction were the sulfotransferase loci both involved in the significance of the<br />

multi-loci HKA test and thus suspected to endure a selective sweep. This is in agreement<br />

with previous findings showing that hybrid populations effectively filter gene flow<br />

between previously isolated lineages via adaptive introgression (Martinsen et al 2001).<br />

However, in this particular case, we cannot rule out the hypothesis that the gene is<br />

evolving very slowly and just un<strong>de</strong>r a parapatric process of differentiation. In<strong>de</strong>ed, since<br />

the sweep_bott software failed to <strong>de</strong>tect a selective sweep, results from the HKA test may<br />

simply reflect strong purifying selection in the divergence between the Pacific and the<br />

Atlantic Bathymodiolus species, Sulfo 1 and 2 genes having virtually no mutation<br />

accumulated in the divergence.<br />

Detection of past hybridization/introgression events between mt cla<strong>de</strong>s using IMa<br />

computation indicated that allele introgression between the two genomes was mostly<br />

orientated from North to South (m1>m2). Again, this finding fits well with the hypothesis<br />

of a moving hybrid zone and/or a <strong>de</strong>nsity-<strong>de</strong>pen<strong>de</strong>nt introgression, in which allele transfer<br />

is more likely to occur in the species having the greatest range expansion (Barton 1986,<br />

Johannesen et al 2006). With the exception of the Lyso gene, allele introgression between<br />

cla<strong>de</strong>s was in the opposite direction when compared to the migration orientation <strong>de</strong>tected<br />

by Migrate_n for all genes and conform to the hypothesis that the spread of newly arriving<br />

alleles is <strong>de</strong>layed by a strong selection against hybrids in the receiving lineage (Nagylaki<br />

1975, Barton 1979, Barton & Hewitt 1985). On the opposite, because the gene Lyso was<br />

the only locus that displayed significantly negative Tajima’s D and Fu & Li tests and<br />

significant LRT for partial selective sweeps (sweep_bott) in both si<strong>de</strong>s of the barrier, one<br />

can propose that an advantageous allele swept in the Northern EPR and subsequently<br />

spread across the barrier in the Southern part of the EPR. Favourable mutations (or alleles<br />

from loci linked to them) can rapidly sweep (or be entrained by genetic hitch-hikking) into<br />

the whole metapopulation when the barrier to gene flow became more permeable (e.g.<br />

Barton 2000, Santiago & Caballero 2005). Such a process was recently shown between the<br />

two hybridizing blue mussels Mytilus edulis and M. galloprovincialis for which genetic<br />

205


Chapitre 4<br />

hich-hikking at a nuclear gene has been <strong>de</strong>monstrated across the Lusitano-Boreal transition<br />

zone (Faure et al 2008). The Lyso gene also displayed the highest mean time of gene flow<br />

(mtime) suggesting that the introgression may be much ol<strong>de</strong>r than the other studied genes<br />

The fact that strong allele recombination between the two divergent cla<strong>de</strong>s have been<br />

<strong>de</strong>tected and especially in the northern part of the EPR (see Fig. 4.17) strengthens this<br />

evolutive scenario since inter-lineage recombination must be old enough to allow newly<br />

recombined alleles to reach a high enough frequency to be sampled.<br />

Testing the ‘Population expansion in the south’ hypothesis from a multiloci dataset<br />

Expansion of the southern populations after the isolation process was already<br />

hypothesized from the multi-species approach because significant negative Tajima’s D<br />

were only found in the Southern lineage of all species (Plouviez et al 2009). Such an<br />

assumption however requires to be also tested using a multigene approach. From our<br />

present dataset, Tajima’s D and Fu & Li F* in<strong>de</strong>xes of nuclear genes are all significantly<br />

negative in the southern cla<strong>de</strong> and, thus in accordance with the expansion hypothesis.<br />

However, while smallest, some genes also displayed significant negative values at these<br />

in<strong>de</strong>xes in the northern cla<strong>de</strong>. A more sophisticated likelihood method was thus used to<br />

discriminate between a selective sweep mo<strong>de</strong>l and a bottleneck mo<strong>de</strong>l (see: Galtier et al<br />

2000). The Sweep_bott test gave interesting information about sweep versus bottleneck<br />

hypotheses. Departures to the mutation/drift equilibrium un<strong>de</strong>r neutrality were <strong>de</strong>tected<br />

both in Northern and Southern cla<strong>de</strong>s and found compatible with either a bottleneck or a<br />

selective sweep mo<strong>de</strong>l. Even if it was not possible to discriminate between a bottleneck<br />

and a sweep in the Northern cla<strong>de</strong>, only one gene (Lyso) displayed a significant <strong>de</strong>parture<br />

to the neutral mo<strong>de</strong>l. Because selection is expected to impact very few genes while<br />

<strong>de</strong>mography affects the whole genome (Hudson et al 1987, Galtier et al 2000), this result<br />

suggests that the population from the North EPR was not subjected to a bottleneck after the<br />

isolation event (ca 1.3 Mya). Such a scenario is in good agreement with Tajima’s and Fu &<br />

Li’s tests and also supports previous works on COI polymorphisms associated with other<br />

vent taxa collected in the same area (Plouviez et al 2009). On the contrary, likelihood ratio<br />

tests in the South indicated the probable occurrence of, at least, two <strong>de</strong>partures to neutral<br />

hypothesis (SAHH and Lyso) and a lack of polymorphism at the Sulfo1 loci (only one<br />

highly frequent allele and a few singletons). With the exception of Sulfo2, all genes<br />

206


Chapitre 4<br />

(including COI) were not following a neutral mo<strong>de</strong>l of mutation accumulation in the<br />

polymorphism with a significant excess of rare variants typifying a situation of an<br />

expanding population. This result fits well with the Isolation with Migration mo<strong>de</strong>l results<br />

suggesting that the Southern cla<strong>de</strong> has a much greater effective population size than those<br />

of the Northern one and the ancestral population: a situation also encountered in the mid-<br />

Atlantic Ridge for the Bathymodiolus mussels (Faure et al 2009).<br />

Present results using nuclear genes thus confirmed the hypothesis of a Southern<br />

<strong>de</strong>mographic expansion in the EPR as previously proposed by Plouviez et al (2009) using a<br />

comparative phylogeography approach among co-distributed species. The greatest tectonic<br />

instability of this region and its subsequent higher extinction rate when compared with the<br />

North province could explain why such a bottleneck was only <strong>de</strong>tected in the South. The<br />

supposedly more probable migration route from South to North (mt haplotypes admixture<br />

in the North) obtained from the multi-loci analysis and the relative strength of the 7°25S-<br />

14°S barrier reinforces this scenario as it seems to prevent the progressive and stepwise<br />

colonisation of the southern EPR by populations from the north: a scenario that would<br />

have also led to an expanding phase.<br />

Acknowledgments<br />

We thank the chief scientists and ‘Nautile’ and ‘Alvin’ crews for their technical support<br />

and efforts during the oceanographic expeditions: HOT96 and BIOSPEEDO2004. We are<br />

very grateful to Claire Daguin-Thiébaut and Frédérique Viard for collecting/preserving<br />

mussels during the BIOSPEEDO cruise. We are also truly in<strong>de</strong>bted to the sequencing<br />

genomic plateforms: GENOMER (<strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>, France) and<br />

GENOSCOPE (Evry, France) for the sequencing of the mtCOI fragment and our clone<br />

collections, respectively. This work was supported by the GDR Ecchis and the ANR-06-<br />

BDV-005 (Deep Oases: coord. D. Desbruyères). We also want to thank A. Tanguy for his<br />

contribution to the B. azoricus cDNA library construction, which provi<strong>de</strong>d exonic<br />

sequences for the nuclear genes and the ‘Bivalvomix’ GIS programme (coord. N. Bierne)<br />

that partly supported the sequencing costs of the library together with the Marine<br />

Genomics Europe NoE. S. Plouviez and B. Faure were supported by a PhD grant from the<br />

Université Pierre et Marie Curie and Marine Genomics Europe NoE (Fish & Shellfish<br />

no<strong>de</strong>), respectively.<br />

207


Chapitre 4<br />

References<br />

Arellano SM, Young CM (2009) Spawning, <strong>de</strong>velopment and the duration of larval life in<br />

a <strong>de</strong>ep-sea cold-seep mussel. Biological Bulletin 216: 149-162.<br />

Avise JC (1994) Molecular markers, Natural History and Evolution. New York: Chapman<br />

& Hall.<br />

Avise JC (2000) Phylogeography: the history and formation of species. Harvard<br />

University Press, Cambridge, MA, USA.<br />

Ban<strong>de</strong>lt HJ, Forster P, Rohl A (1999) Median-joining networks for inferring intraspecific<br />

phylogenies. Molecular Biology and Evolution 16, 37-48.<br />

Barton NH (1979) Gene Flow Past a Cline. Heredity 43: 333-339.<br />

Barton NH, Hewitt GM (1985) Analysis of Hybrid Zones. Annual Review of Ecology and<br />

Systematics 16, 113-148.<br />

Barton NH (1986) The Effects of Linkage and Density-Depen<strong>de</strong>nt Regulation on Gene<br />

Flow. Heredity 57, 415-426.<br />

Barton NH, Hewitt GM (1989) Adaptation, speciation and hybrid zones. Nature 341: 497-<br />

503.<br />

Barton NH (2000) Estimating multilocus linkage disequilibria. Heredity 84, 373-389.<br />

Beerli P, Felsenstein J (1999) Maximum-likelihood estimation of migration rates and<br />

effective population numbers in two populations using a coalescent approach.<br />

Genetics 152: 763-773.<br />

Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and<br />

effective population sizes in n subpopulations by using a coalescent approach.<br />

Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of America<br />

98: 4563-4568.<br />

Belkhir K, Borsa P, Chikhi L, Raufaste N, Bonhomme F (2004) Genetix 4.05, logiciel sous<br />

Windows TM <strong>pour</strong> la génétique <strong>de</strong>s populations. Laboratoire Génome, Populations,<br />

Interactions, Adaptations, UMR 5000, Université <strong>de</strong> Montpellier 2. Available at:<br />

www.genetix.univ-montp2.fr/ genetix/genetix.htm<br />

Bierne N, Tanguy A, Faure M, Faure B, David E, et al (2007) Mark-recapture cloning: a<br />

straightforward and cost-effective cloning method for population genetics of singlecopy<br />

nuclear DNA sequences in diploids. Molecular Ecology Notes 7: 562-566.<br />

Black MB, Lutz RA, Vrijenhoek RC (1994) Gene Flow among Vestimentiferan Tube<br />

Worm (Riftia-Pachyptila) Populations from Hydrothermal Vents of the Eastern<br />

Pacific. Marine Biology 120, 33-39.<br />

Boutet I, Tanguy A, Le Guen D, Piccino P, Hour<strong>de</strong>z S, Legendre P, Jollivet D (2009)<br />

Global <strong>de</strong>pression in gene expression as a response to rapid thermal changes in vent<br />

mussels. Proceedings of the Royal Society of London Series B.<br />

doi:10.1098/rspb.2009.0503<br />

Cannon GA, Pashinski DJ (1997) Variation in mean currents affecting hydrothermal<br />

plumes of the Juan <strong>de</strong> Fuca Ridge. Journal of Geophysical Research 102: 24965-<br />

24976.<br />

Carney SL, Formica MI, Divatia H, Nelson K, Fisher CR, Schaeffer SW (2006) Population<br />

structure of the mussel “Bathymodiolus” childressi from Gulf of Mexico hydrocarbon<br />

seeps. Deep-Sea Research I 53: 1061-1072.<br />

Chevaldonné P, Jollivet D, Vangriesheim A, Desbruyeres D (1997) Hydrothermal-vent<br />

alvinellid polychaete dispersal in the eastern Pacific .1. Influence of vent site<br />

distribution, bottom currents, and biological patterns. Limnology and Oceanography<br />

42, 67-80.<br />

208


Chapitre 4<br />

Cox BC (2001) The biogeographic regions reconsi<strong>de</strong>red. Journal of Biogeography 28,<br />

511-523.<br />

Craddock C, Hoeh WR, Lutz RA, Vrijenhoek RC (1995) Extensive Gene Flow among<br />

Mytilid (Bathymodiolus thermophilus) Populations from Hydrothermal Vents of the<br />

Eastern Pacific. Marine Biology 124, 137-146.<br />

Craddock C, Lutz RA, Vrijenhoek RC (1997) Patterns of dispersal and larval <strong>de</strong>velopment<br />

of archaeogastropod limpets at hydrothermal vents in the eastern Pacific. Journal of<br />

Experimental Marine Biology and Ecology 210, 37-51.<br />

Dawson MN (2001) Phylogeography in coastal marine animals: a solution from<br />

California? Journal of Biogeography 28, 723-736.<br />

Faure B, Jollivet D, Tanguy A, Bonhomme F, Bierne N (2009) Speciation in the <strong>de</strong>ep-sea:<br />

multilocus analysis of divergence and gene flow between two hybridizing species of<br />

hydrothermal vent mussels. PlosOne.<br />

Faure MF, David P, Bonhomme F, Bierne N (2008) Genetic hitchhiking in a subdivi<strong>de</strong>d<br />

population of Mytilus edulis. Bmc Evolutionary Biology 8.<br />

France SC, Hessler RR, Vrijenhoek RC (1992) Genetic differentiation between spatiallydisjunct<br />

populations of the <strong>de</strong>ep-sea, hydrothermal vent-en<strong>de</strong>mic amphipod Ventiella<br />

sulfuris. Marine Biology 114, 551-559.<br />

Francheteau J, Armijo R, Cheminee JL, et al (1990) 1 Ma East Pacific Rise Oceanic-Crust<br />

and Uppermost Mantle Exposed by Rifting in Hess Deep (Equatorial Pacific-Ocean).<br />

Earth and Planetary Science Letters 101: 281-295.<br />

Fu YX, Li WH (1993) Statistical test of neutrality of mutations. Genetics 133: 693-709.<br />

Fujio SZ, Imasato N (1991) Diagnostic Calculation for Circulation and Water Mass<br />

Movement in the Deep Pacific. Journal of Geophysical Research-Oceans 96, 759-774.<br />

Fustec A, Desbruyères D, Juniper SK (1987) Genetic differentiation between spatiallydisjunct<br />

populations of the <strong>de</strong>ep-sea, hydrothermal vent-en<strong>de</strong>mic amphipod Ventiella<br />

sulfuris. Biological Oceanography 4, 121-164.<br />

Galtier N, Depaulis F, Barton NH (2000) Detecting bottlenecks and selective sweeps from<br />

DNA sequence polymorphism. Genetics 155, 981-987.<br />

Grassle JP (1985) Genetic differentiation in populations of hydrothermal vent mussels<br />

(Bathymodiolus thermophilus) from the Galapagos Rift and 13°N on the East Pacific<br />

Rise. Biological Society of Washington 6, 429-442.<br />

Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis<br />

program for Windows 95/98/NT. Nucleic Acids Symposium Series 41, 95-98.<br />

Harrison RG (1990) Hybrid zones: windows on evolutionary processes. Oxf. Surv. Evol.<br />

Biol. 7: 69-128.<br />

Hasegawa M, Kishino H, Yano T (1985) Dating the human-ape splitting by a molecular<br />

clock of mitochondrial DNA. Journal of Molecular Evolution 22: 160-174.<br />

Haymon RM, Fornari DJ, Edwards MH, et al. (1991) Hydrothermal Vent Distribution<br />

Along the East Pacific Rise Crest (9-Degrees-09'-54'n) and Its Relationship to<br />

Magmatic and Tectonic Processes on Fast-Spreading Midocean Ridges. Earth and<br />

Planetary Science Letters 104, 513-534.<br />

Hey J, Nielsen R (2004) Multilocus methods for estimating population sizes, migration<br />

rates and divergence time, with applications to the divergence of Drosophila<br />

pseudoobscura and D. persimilis. Genetics 167, 747-760.<br />

Hudson RR, Kaplan NL (1985) Statistical Properties of the Number of Recombination<br />

Events in the History of a Sample of DNA-Sequences. Genetics 111, 147-164.<br />

Hudson RR, Kreitman M, Aguadé M (1987) A test of neutral molecular evolution based on<br />

nucleoti<strong>de</strong> data. Genetics 116: 153-159.<br />

209


Chapitre 4<br />

Hudson RR, Slatkin M, Maddison WP (1992) Estimation of Levels of Gene Flow from<br />

DNA-Sequence Data. Genetics 132, 583-589.<br />

Hurtado L, Lutz R, Vrijenhoek R (2004) Distinct patterns of genetic differentiation among<br />

annelids of eastern Pacific hydrothermal vents. Molecular Ecology 13, 2603-2615.<br />

Johannesen J, Johannesen B, Griebeler EM, et al. (2006) Distortion of symmetrical<br />

introgression in a hybrid zone: evi<strong>de</strong>nce for locus-specific selection and unidirectional<br />

range expansion. Journal of Evolutionary Biology 19, 705-716.<br />

Jollivet D, Desbruyeres D, Bonhomme F, Moraga D (1995) Genetic Differentiation of<br />

Deep-Sea Hydrothermal Vent Alvinellid Populations (Annelida, Polychaeta) Along<br />

the East Pacific Rise. Heredity 74, 376-391.<br />

Jollivet D, Chevaldonne P, Planque B (1999) Hydrothermal-vent alvinellid polychaete<br />

dispersal in the eastern Pacific. 2. A metapopulation mo<strong>de</strong>l based on habitat shifts.<br />

Evolution 53, 1128-1142.<br />

Jollivet D, Empis A, Baker MC, et al. (2000) Reproductive biology, sexual dimorphism,<br />

and population structure of the <strong>de</strong>ep sea hydrothermal vent scale-worm,<br />

Branchipolynoe seepensis (Polychaeta : Polynoidae). Journal of the Marine Biological<br />

Association of the United Kingdom 80, 55-68.<br />

Jolly MT, Viard F, Gentil F, Thiebaut E, Jollivet D (2006) Comparative phylogeography of<br />

two coastal polychaete tubeworms in the Northeast Atlantic supports shared history<br />

and vicariant events. Molecular Ecology 15, 1841-1855.<br />

Kervévan (1991) Modélisation tridimentionnelle <strong>de</strong> la convection hydrothermale en zone<br />

<strong>de</strong> rift océanique. Ph.D. dissertation, Université <strong>de</strong> Bretagne Occi<strong>de</strong>ntale, Brest,<br />

France.<br />

Kureth CL, Rea DK (1981) Large-Scale Oblique Features in an Active Transform-Fault,<br />

the Wilkes Fracture-Zone near 9-Degrees-S on the East Pacific Rise. Marine<br />

Geophysical Researches 5: 119-137.<br />

Le Pennec M, Hily A, Lucas A (1984) Structures gonadiques particulières d’un Mytilidae<br />

profound <strong>de</strong>s sources hydrothermales du Pacifique oriental. Comptes Rendus <strong>de</strong><br />

l’Académie <strong>de</strong>s Sciences <strong>de</strong> Paris, Série III 299, 725-730.<br />

Lupton JE, Craig H (1981) A Major He-3 Source at 15-Degrees-S on the East Pacific Rise.<br />

Science 214, 13-18.<br />

Lupton J (1998) Hydrothermal helium plumes in the Pacific Ocean. Journal of<br />

Geophysical Research-Oceans 103, 15853-15868.<br />

Lyrholm T, Leimar O, Johanneson B, Gyllensten U (1999) Sex-biased dispersal in sperm<br />

whales: contrasting mitochondrial and nuclear genetic structure of global populations.<br />

Proceeding of the Royal Society of London Series B 266: 347-354.<br />

Maggs CA, Castilho R, Foltz D, Henzler C, Jolly MT, Kelly J, Olsen J, Perez KE, Stam W,<br />

Väinölä R, Viard F, Wares J (2008) Evaluating signatures of glacial refugia for north<br />

atlantic benthic marine taxa. Ecology 89: 108-122.<br />

Marcus J, Tunnicliffe V, Butterfield DA (2009) Post-eruption succession of macrofaunal<br />

communities at diffuse flow hydrothermal vents on Axial Volcano, Juan <strong>de</strong> Fuca<br />

Ridge, Northeast Pacific. Deep-Sea Research Part Ii-Topical Studies in Oceanography<br />

56, 1586-1598.<br />

Marsh AG, Mullineaux LS, Young CM, Manahan DT (2001) Larval dispersal potential of<br />

the tubeworm Riftia pachyptila at <strong>de</strong>ep-sea hydrothermal vents. Nature 411, 77-80.<br />

Martinsen GD, Whitham TG, Turek RJ, Keim P (2001) Hybrid populations selectively<br />

filter gene introgression between species. Evolution 55, 1325-1335.<br />

McDonald JH, Kreitman M (1991) Adaptive Protein Evolution at the Adh Locus in<br />

Drosophila. Nature 351, 652-654.<br />

210


Chapitre 4<br />

Mullineaux LS, Mills SW, Sweetman AK, Beaudreau AH, Metaxas A, Hunt HL (2005)<br />

Vertical, lateral and temporal structure in larval distributions at hydrothermal vents.<br />

Marine Ecology Progress Series 293, 1-16.<br />

Naar DF, Hey RN (1989) Speed Limit for Oceanic Transform Faults. Geology 17: 420-<br />

422.<br />

Nagylaki T (1975) Conditions for the existence of clines. Genetics 80: 595-615.<br />

Nielsen R, Wakeley J (2001) Distinguishing migration from isolation: A Markov chain<br />

Monte Carlo approach. Genetics 158, 885-896.<br />

Olu-Le Roy L, von Cosel R, Hour<strong>de</strong>z S, Carney SL, Jollivet D (2007) Amphi-Atlantic<br />

cold-seep Bathymodiolus species complexes across the equatorial belt. Deep-Sea<br />

Research I 54, 1890-1911.<br />

Pasteur N, Pasteur G, Bonhomme F, Catalan J, Britton-Davidian J (1987) Manuel<br />

technique <strong>de</strong> génétique par électrophorèse <strong>de</strong> protéines. Paris: Lavoisier.<br />

Piccino P, Viard F, Sarradin PM, et al (2004) Thermal selection of PGM allozymes in<br />

newly foun<strong>de</strong>d populations of the thermotolerant vent polychaete Alvinella<br />

pompejana. Proceedings of the Royal Society of London Series B-Biological Sciences<br />

271, 2351-2359.<br />

Plouviez S, Shank TM, Faure B, Daguin-Thiébaut C, Viard F, Lallier FH, Jollivet D<br />

(2009) Comparative phylogeography among hydrothermal vent species along the East<br />

Pacific Rise reveals vicariant processes and population expansion in the South.<br />

Molecular Ecology 18: 3903-3917.<br />

Rieseberg LH, Whitton J, Gardner K (1999) Hybrid zones and the genetic architecture of a<br />

barrier to gene flow between two sunflower species. Genetics 152: 713-727.<br />

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA<br />

polymorphism analyses by the coalescent and other methods. Bioinformatics 19, 2496-<br />

2497.<br />

Santiago E, Caballero A (2005) Variation after a selective sweep in a subdivi<strong>de</strong>d<br />

population. Genetics 169, 475-483.<br />

Shank TM, Fornari DJ, Von Damm KL et al (1998) Temporal and spatial patterns of<br />

biological community <strong>de</strong>velopment at nascent <strong>de</strong>ep-sea hydrothermal vents (9°50’N,<br />

East Pacific Rise). Deep Sea Research Part II: Topical Studies in Oceanography 45,<br />

465-515.<br />

Tajima F (1989) Statistical-method for testing the neutral mutation hypothesis by DNA<br />

polymorphism. Genetics 123, 585-595.<br />

Thiel M, Macaya EC, Acuna E, et al. (2007) The Humboldt Current System of northern<br />

and central Chile. Oceanography and Marine Biology, Vol 45 45, 195-344.<br />

Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of<br />

progressive multiple sequence alignment through sequence weighting, position<br />

specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673-<br />

4680.<br />

Thomson RE, Roth SE, Dymond J (1990) Near-inertial motion over a mid-ocean ridge:<br />

effects of topography and hydrothermal plumes. Journal of Geophysical Research 95:<br />

12961-12966.<br />

Trewick SA, Wallis GP, Morgan-Richards M (2000) Phylogeographical pattern correlates<br />

with Pliocene mountain building in the alpine scree weta (Orthoptera,<br />

Anostostomatidae). Molecular Ecology 9, 657-666.<br />

Tunnicliffe V, Embley RW, Hol<strong>de</strong>n JF, et al. (1997) Biological colonization of new<br />

hydrothermal vents following an eruption on Juan <strong>de</strong> Fuca Ridge. Deep-Sea Research<br />

Part I-Oceanographic Research Papers 44, 1627-&.<br />

211


Chapitre 4<br />

Tyler PA, Young CM (1999) Reproduction and dispersal at vents and cold seeps. Journal<br />

of the Marine Biological Association of the United Kingdom 79, 193-208.<br />

Vrijenhoek RC (1997) Gene flow and genetic diversity in naturally fragmented<br />

metapopulations of <strong>de</strong>ep-sea hydrothermal vent animals. Journal of Heredity 88, 285-<br />

293.<br />

Watremez P, Kervévan C (1990) Origin of Variations in Hydrothermal Activity - 1st<br />

Results of a Simple Numerical-Mo<strong>de</strong>l. Comptes Rendus De L Aca<strong>de</strong>mie Des Sciences<br />

Serie Ii 311, 153-158.<br />

Weir B, Cockerham C (1984) Estimating F-statistics for the analysis of population<br />

structure. Evolution 38: 1358-1370.<br />

Won Y, Young CR, Lutz RA, Vrijenhoek RC (2003) Dispersal barriers and isolation<br />

among <strong>de</strong>ep-sea mussel populations (Mytilidae : Bathymodiolus) from eastern Pacific<br />

hydrothermal vents. Molecular Ecology 12, 169-184.<br />

212


Chapitre 4<br />

Supplementary material.<br />

S1: Allele frequencies of enzyme loci for B. thermophilus populations of each vent<br />

field of the SEPR.N, population size; H exp, expected heterozygosity; H obs, observed<br />

heterozygosity.<br />

7°25’S 14°S 17°25’S 17°35’S 18°33’S 21°33’S<br />

Pgm<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

64 0 0 0.0036 0 0 0<br />

74 0.05 0.0132 0.0109 0 0.0455 0.0194<br />

81 0.7 0.5789 0.5219 0.5652 0.5758 0.5194<br />

85 0 0 0.0036 0 0 0.0146<br />

89 0.25 0.3816 0.4307 0.3696 0.3636 0.4223<br />

95 0 0.0263 0.0292 0.0652 0.0152 0.0243<br />

H exp. 0.445 0.5184 0.5412 0.5397 0.534 0.5507<br />

H obs. 0.5 0.5789 0.5255 0.6087 0.3636 0.4272<br />

Mpi<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

100 0.05 0.0395 0.0182 0.0435 0 0<br />

120 0.95 0.9474 0.9781 0.9348 1 0.9903<br />

140 0 0.0132 0 0.0217 0 0.0097<br />

160 0 0 0.0036 0 0 0<br />

H exp. 0.095 0.1008 0.043 0.1238 0 0.0192<br />

H obs. 0.1 0.1053 0.0438 0.087 0 0.0194<br />

Odh<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

60 0 0.0658 0.0073 0 0.0303 0.0097<br />

80 1 0.9342 0.9526 1 0.9545 0.9612<br />

100 0 0 0.0401 0 0.0152 0.0291<br />

H exp. 0 0.1229 0.091 0 0.0877 0.0752<br />

H obs. 0 0.1316 0.0949 0 0.0909 0.0777<br />

Lap<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

75 0 0.0132 0 0 0 0.0049<br />

82 0.9 0.7632 0.719 0.7609 0.7879 0.7233<br />

90 0 0.0132 0.0073 0 0.0303 0.0097<br />

100 0.1 0.2105 0.2482 0.2174 0.1364 0.233<br />

110 0 0 0.0255 0.0217 0.0455 0.0291<br />

H exp. 0.18 0.3729 0.4208 0.3733 0.3577 0.4216<br />

H obs. 0.2 0.3684 0.4088 0.3043 0.303 0.4078<br />

213


Chapitre 4<br />

7°25’S 14°S 17°25’S 17°35’S 18°33’S 21°33’S<br />

Gpi<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

88 0 0.0132 0.0219 0.0435 0 0<br />

100 0 0 0 0 0 0.0049<br />

112 0.85 0.9605 0.938 0.8913 0.9545 0.9466<br />

116 0.05 0 0 0 0 0<br />

120 0.1 0.0263 0.0401 0.0652 0.0455 0.0485<br />

H exp. 0.265 0.0765 0.1181 0.1994 0.0868 0.1016<br />

H obs. 0.3 0.0789 0.1095 0.2174 0.0909 0.1068<br />

Mdh low<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

74 1 1 0.9927 1 1 1<br />

100 0 0 0.0073 0 0 0<br />

H exp. 0 0 0.0145 0 0 0<br />

H obs. 0 0 0.0146 0 0 0<br />

Mdh high<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

100 0 0.1711 0.073 0.0217 0.2727 0.1019<br />

120 1 0.8289 0.927 0.9783 0.7273 0.8981<br />

H exp. 0 0.2836 0.1353 0.0425 0.3967 0.1831<br />

H obs. 0 0.3421 0.146 0.0435 0.5455 0.2039<br />

Hk<br />

(N) 10 38 137 23 33 103<br />

Allele<br />

70 0.05 0 0.0036 0.0217 0 0<br />

100 0.9 1 0.9964 0.9783 1 1<br />

130 0.05 0 0 0 0 0<br />

H exp. 0.185 0 0.0073 0.0425 0 0<br />

H obs. 0.2 0 0.0073 0.0435 0 0<br />

214


Chapitre 4<br />

S2 : Observed <strong>de</strong>lta theta of studied genes in the distribution simulated un<strong>de</strong>r neutral<br />

hypothesis<br />

215


Chapitre 4<br />

4.3. L. elevatus<br />

4.3.1. Introduction<br />

Le complexe d’espèces L. elevatus situé sur l’EPR, compte <strong>de</strong>ux espèces<br />

majoritairement abondantes divergentes <strong>de</strong> 6,5% sur le gène mtCOI (Matabos et al<br />

2008b) : L. elevatus McLean (appelée espèce Nord dans notre étu<strong>de</strong>) distribuée <strong>de</strong> 21°N à<br />

9°50’N et une secon<strong>de</strong> espèce (que nous appellerons espèce Sud) <strong>de</strong> 9°50’N à 23°S<br />

(Johnson et al 2008, Plouviez et al 2009). Le site 9°50’N constitue donc une zone <strong>de</strong><br />

sympatrie entre ces <strong>de</strong>ux espèces cryptiques, mais l’espèce Sud y est majoritaire. Cette<br />

distribution géographique suggère une spéciation allopatrique <strong>de</strong> ces <strong>de</strong>ux espèces, datée à<br />

environ 11,6 Ma et probablement située entre 13°N et 9°50’N, coïncidant temporellement<br />

et spatialement avec la réorientation <strong>de</strong> la ri<strong>de</strong> Mathématicienne (Mammerickx & Klitgord<br />

1982, Plouviez et al 2009). Le processus <strong>de</strong> spéciation a pu se renforcer par une spéciation<br />

écologique, l’espèce Nord étant trouvée principalement en association avec les<br />

siboglinidae Riftia pachyptila et l’espèce Sud avec les moules hydrothermales B.<br />

thermophilus. Néanmoins, la séparation <strong>de</strong>s 2 espèces n’est pas complète, <strong>de</strong>s individus<br />

hybri<strong>de</strong>s ayant été détectées à 9°50’N à partir d’un génotypage allozymique (Matabos et al<br />

2008b). Le statut d’hybri<strong>de</strong> entre ces espèces cryptiques n’a <strong>pour</strong> l’instant pas été évalué<br />

<strong>de</strong> façon certaine car les individus séquencés sur l’ADN mitochondrial ne correspondaient<br />

pas aux individus génotypés en allozymes. Une telle approche mériterait d’être effectuée<br />

en élargissant le champ d’investigation géographique à l’ensemble <strong>de</strong> l’EPR. De plus, le<br />

second événement <strong>de</strong> vicariance plus récent (1,3 Ma) observé <strong>pour</strong> toutes les espèces entre<br />

les populations sud et nord <strong>de</strong> l’EPR semble avoir également impacté l’espèce Sud <strong>de</strong> L.<br />

elevatus à l’équateur, en séparant la population <strong>de</strong> 9°50’N <strong>de</strong>s autres populations situées<br />

plus au sud. Cet effet est observable en comparant les séquences du gène mtCOI entre<br />

individus <strong>de</strong> la lignée sud. La mise en place d’une barrière secondairement à la séparation<br />

<strong>de</strong>s espèces <strong>de</strong> L. elevatus <strong>pour</strong>rait jouer un rôle fondamental dans la dynamique<br />

d’hybridation <strong>de</strong>s espèces, notamment en termes <strong>de</strong> longueur <strong>de</strong> la zone d’introgression et<br />

nécessite une attention particulière.<br />

Nous avons ainsi utilisé, en plus du gène mtCOI, trois gènes nucléaires séquencés<br />

sur <strong>de</strong>s individus répartis <strong>de</strong> 21°N à 21°S, afin <strong>de</strong> mieux comprendre le rôle <strong>de</strong>s barrières<br />

216


Chapitre 4<br />

(physique et génétique) 13°N-9°50N et 9°50’N-7°25S dans la structuration du complexe<br />

d'espèces et la dynamique d'hybridation.<br />

4.3.2. Gène mitochondrial COI<br />

4.3.2.1. Doubles pics sur les chromatogrammes mtCOI <strong>de</strong> certains<br />

individus<br />

Sur l’ensemble <strong>de</strong>s individus séquencés le long <strong>de</strong> l’EPR, six individus appartenant<br />

à la zone <strong>de</strong> contact entre les 2 espèces cryptiques (9°50’N) présentaient <strong>de</strong>s profils <strong>de</strong><br />

séquences COI atypiques. En effet, si les autres individus présentaient <strong>de</strong>s<br />

chromatogrammes clairs correspondant à l’une ou l’autre <strong>de</strong>s lignées mtCOI, les<br />

chromatogrammes <strong>de</strong> ces 6 individus (séquencés en direct) comptaient <strong>de</strong>s doubles pics<br />

suggérant l’amplification <strong>de</strong> 2 haplotypes (Fig. 4.22). Bien que <strong>de</strong>s exceptions existent<br />

(e.g. Skibinski et al 1994 ; Zouros et al 1994), le génome mitochondrial est théoriquement<br />

hérité uniquement par la voie maternelle et ne <strong>de</strong>vrait donc pas présenter <strong>de</strong> tels profils. De<br />

plus, les doubles pics observés correspon<strong>de</strong>nt systématiquement aux pics attendus <strong>pour</strong> les<br />

mutations fixées <strong>de</strong> chacune <strong>de</strong>s 2 espèces. Ainsi, ces individus présenteraient à la fois<br />

l’haplotype <strong>de</strong> l’espèce nord et l’haplotype <strong>de</strong> l’espèce sud.<br />

217


Chapitre 4<br />

Fig. 4.22 Portion <strong>de</strong> chromatogramme présentant <strong>de</strong>s doubles pics correspondant<br />

respectivement à une base fixée chez l’espèce Nord (in<strong>de</strong>xée en bleu sous le<br />

chromatogramme) et une base fixée chez l’espèce Sud (in<strong>de</strong>xée en rouge).<br />

Pour nous assurer <strong>de</strong> la présence <strong>de</strong> ces 2 types <strong>de</strong> mitochondrie et vali<strong>de</strong>r<br />

l’hypothèse d’hétéroplasmie chez ces 6 individus ‘hybri<strong>de</strong>s’, nous avons cloné leur produit<br />

<strong>de</strong> PCR (clonage individuel). Les clones repiqués (8 par individus lorsque les conditions le<br />

permettaient) sont ensuite amplifiés via <strong>de</strong>s amorces plasmidiques selon le protocole décrit<br />

en annexe 3. Une digestion enzymatique <strong>de</strong> l’ADN (1 µl d’enzyme <strong>de</strong> restriction Sac1 à<br />

une concentration <strong>de</strong> 20U/µl, 2,5 µl <strong>de</strong> tampon 10X associé à cette enzyme, 5 µl <strong>de</strong> produit<br />

<strong>de</strong> PCR, réalisée sur les clones repiqués, et 12,5 µl d’H 2 0, 3h à 37°C) permettant <strong>de</strong><br />

générer <strong>de</strong>s patrons RFLP discriminant les mitochondries <strong>de</strong> type nord et sud, a ensuite été<br />

réalisée sur ces clones repiqués. Ces individus présentaient bien les 2 types <strong>de</strong> profils <strong>de</strong><br />

RFLP (Fig. 4.23).<br />

Exemples <strong>de</strong> profils<br />

mitochondrial <strong>de</strong> type<br />

espèce Nord<br />

Exemples <strong>de</strong> profils<br />

mitochondrial <strong>de</strong> type<br />

espèce Sud<br />

218<br />

Fig. 4.23 Visualisation sur gel d’agarose <strong>de</strong>s <strong>de</strong>ux types <strong>de</strong> profil RFLP sur les 8<br />

clones d’un individu présentant les <strong>de</strong>ux types <strong>de</strong> mitochondrie.<br />

Chaque puit correspond un profil RFLP réalisé sur un clone repiqué.


Chapitre 4<br />

Nous avons séquencé alors les clones correspondant à ces 2 profils RFLP, vérifiant<br />

ainsi qu’ils correspondaient bien aux profils mitochondriaux <strong>de</strong>s 2 espèces cryptiques. Les<br />

2 types <strong>de</strong> mitochondries sont donc bien amplifiés <strong>pour</strong> ces 6 individus échantillonnés à<br />

9°50’N à chromatogramme ambiguë. Bien que les témoins négatifs <strong>de</strong> PCR soient sans<br />

ambiguïté, une secon<strong>de</strong> PCR et un second séquençage en direct <strong>de</strong> ces individus ont été<br />

réalisés afin d’éliminer <strong>de</strong> possibles contaminations <strong>de</strong> produits d’amplification (lors <strong>de</strong> la<br />

purification). Les mêmes résultats ont été obtenus (remarque : ces individus n’ont pas été<br />

utilisés dans Plouviez et al 2009).<br />

4.3.2.2. Hypothèse <strong>de</strong> double hérédité mitochondriale<br />

En général, l’ADN mitochondrial <strong>de</strong>s animaux est exclusivement hérité<br />

maternellement (Birky 2001) et toutes les cellules d’un même organisme présentent la<br />

même copie <strong>de</strong> cet ADN, hormis quelques rares mutations (homoplasmie). Cependant,<br />

certaines espèces comptent parfois plusieurs ADN mitochondriaux au sein d’une même<br />

cellule : c’est l’hétéroplasmie. L’ensemble <strong>de</strong>s ADN mitochondriaux au sein d’une cellule<br />

peut alors être considéré comme une population. Un ADN mitochondrial mutant <strong>pour</strong>ra<br />

ainsi, par dérive, transmettre son ADN à une cellule fille (issue <strong>de</strong> la réplication <strong>de</strong> la<br />

cellule initiale) et donc, si cette mutation est délétère, diminuer la valeur sélective <strong>de</strong> cette<br />

cellule (moins <strong>de</strong> mitochondries fonctionnelles, Breton et al 2007). L’ADN mutant entrera<br />

ainsi en compétition, lors <strong>de</strong> sa réplication, avec les ADN non-mutants. Des processus <strong>de</strong><br />

sélection contre les mutants mitochondriaux incapables d’effectuer la phosphorylation<br />

oxydative ont ainsi été mis en évi<strong>de</strong>nce chez la levure (Reid 1980, Birky 1995). L’hérédité<br />

parentale unique est un mécanisme qui est supposé avoir été favorisé car elle permettrait<br />

d’éviter ces conflits intracellulaires (Rand 2001). Cependant, certaines espèces présentent<br />

une double hérédité mitochondriale, notamment les moules marines <strong>de</strong> l’ordre <strong>de</strong>s<br />

Mytiloida, les moules d’eau douces <strong>de</strong> la superfamille <strong>de</strong>s Unionoi<strong>de</strong>a ou les palour<strong>de</strong>s <strong>de</strong><br />

l’ordre <strong>de</strong>s Veneroida (e.g. Skibinski et al 1994, Liu et al 1996, Passamonti & Scali 2001).<br />

La présence <strong>de</strong> <strong>de</strong>ux mtCOI appartenant chacune à une espèce cryptique différente semble<br />

donc indiquer que L. elevatus peut également, au moins chez certains hybri<strong>de</strong>s,<br />

probablement la génération F1, présenter une double hérédité parentale.<br />

219


Chapitre 4<br />

L’hybridation entre espèces différentes peut avoir <strong>de</strong>s conséquences sur l’hérédité<br />

parentale (Wood et al 2003). En effet, chez Mytilis edulis et M. galloprovincialis, la double<br />

hérédité maternelle est généralement présente chez les mâles adultes alors que la<br />

mitochondrie mâle est éliminée chez les femelles (même si elles peuvent subsister dans<br />

certains tissus <strong>de</strong>s femelles, Dalziel et al 2002, Breton et al 2007). Cependant, les mâles<br />

hybri<strong>de</strong>s entre ces <strong>de</strong>ux espèces, avec une mitochondrie mâle provenant <strong>de</strong> M. edulis et<br />

une mitochondrie femelle <strong>de</strong> M. galloprovincialis, élimineront plus aisément la<br />

mitochondrie mâle qu’un croisement issu <strong>de</strong> lignées pures (Wood et al 2003). Au contraire,<br />

les mâles hybri<strong>de</strong>s ayant une mitochondrie mâle issue <strong>de</strong> M. galloprovincialis et une<br />

mitochondrie femelle <strong>de</strong> M. edulis conserveront plus facilement la mitochondrie mâle que<br />

chez les M. galloprovincialis non hybri<strong>de</strong>s (Wood et al 2003). Les chromatogrammes <strong>de</strong>s<br />

L. elevatus non hybri<strong>de</strong>s ne présentant pas <strong>de</strong> double pic apparent, nous pouvons ainsi<br />

supposer que seuls les hybri<strong>de</strong>s F1 possè<strong>de</strong>raient la double hérédité parentale, la<br />

mitochondrie mâle étant très probablement éliminée chez les individus non-hybri<strong>de</strong>s lors<br />

<strong>de</strong> la ségrégation méiotique (à l’instar <strong>de</strong>s femelles Mytilidae, Cao et al 2004) à la fois<br />

chez les femelles et chez les mâles. Bien que nécessitant <strong>de</strong> plus amples étu<strong>de</strong>s,<br />

l’hybridation entre les <strong>de</strong>ux espèces cryptiques <strong>pour</strong>rait avoir favoriser le maintien <strong>de</strong> la<br />

lignée mitochondriale mâle chez certains hybri<strong>de</strong>s, comme chez les M. galloprovincialis<br />

mâles hybri<strong>de</strong>s (Wood et al 2003). Ceci reste néanmoins une hypothèse à tester.<br />

4.3.3. Analyse <strong>de</strong>s séquences nucléaires : hybridation et barrière(s) au(x) flux<br />

<strong>de</strong> gènes<br />

4.3.3.1. Recapture <strong>de</strong> séquences nucléaires<br />

Le nombre d’individus recapturés <strong>pour</strong> chaque gène nucléaire est présenté dans le<br />

tableau 4.14 (chaque individu ayant été séquencé sur le mtCOI). Le faible nombre <strong>de</strong><br />

séquences recapturées <strong>pour</strong> le gène supposé unique <strong>de</strong> la Textilinine résulte <strong>de</strong> la<br />

suppression <strong>de</strong> nombreuses séquences correspondant au(x) gène(s) paralogues coamplifié(s)<br />

avec nos amorces (cf. Chapitre 2). De plus, nous n’avons pas réussi à amplifier<br />

le fragment <strong>de</strong> ce gène chez les individus <strong>de</strong> la population 21°N. Il en résulte que seuls 10<br />

individus ont été recapturés sur les trois gènes nucléaires, mais 60 ont été recapturés sur<br />

<strong>de</strong>ux <strong>de</strong> ces gènes.<br />

220


Chapitre 4<br />

Tableau 4.14 Nombre d’individus recapturés aux différents sites hydrothermaux.<br />

N 3gènes et N 2gènes représentent le nombre d’individus recapturés sur les 3 gènes nucléaires<br />

et sur 2 <strong>de</strong>s gènes, respectivement.<br />

Localité Lysozyme VDG3 Textilinine N 3gènes N 2gènes<br />

21°N 11 8 0 0 6<br />

13°N 18 16 9 1 12<br />

9°50’N 20 23 10 6 14<br />

14°S 11 13 1 1 8<br />

17°25’S 11 11 3 0 8<br />

17°35’S 12 6 0 0 4<br />

21°33’S 19 10 5 2 8<br />

4.3.3.2. Polymorphisme et divergence <strong>de</strong> séquences <strong>pour</strong> les 3 gènes<br />

nucléaires étudiés<br />

Le clonage ne permettant pas <strong>de</strong> recapturer les <strong>de</strong>ux allèles <strong>de</strong> chaque individu, <strong>pour</strong><br />

limiter les biais dans les fréquences alléliques et simplifier la représentation <strong>de</strong>s arbres<br />

phylogénétiques, seule une séquence par individu a été utilisée <strong>pour</strong> l’ensemble <strong>de</strong>s<br />

analyses (jeux <strong>de</strong> données <strong>de</strong> qualité 3). Pour chaque gène, un arbre phylogénétique a été<br />

construit selon la métho<strong>de</strong> d’agglomération <strong>de</strong>s plus proches voisins (Neighbour Joining)<br />

via le logiciel Mega 4.0 (Tamura et al 2007). Pour les gènes Lysozyme et VDG3, la<br />

distribution <strong>de</strong>s allèles a été réalisée au sein <strong>de</strong> chaque population selon leur cla<strong>de</strong><br />

d’appartenance. La distribution allélique selon le cla<strong>de</strong> d’appartenance n’a pas pu être<br />

réalisée sur le gène Textilinine en raison du faible nombre d’individus recapturés au sein<br />

<strong>de</strong> chaque site (généralement inférieur à 10 individus par site, Tableau 4.14).<br />

Les allèles du gène Lysozyme (Fig. 4.24.A) peuvent être divisés en trois cla<strong>de</strong>s : le<br />

cla<strong>de</strong> 1 comptant l’ensemble <strong>de</strong>s séquences <strong>de</strong>s sites 21°N et 13°N (espèce Nord) tandis<br />

que les séquences <strong>de</strong>s individus provenant <strong>de</strong>s sites 14°S à 21°33’S, sont distribuées dans<br />

les cla<strong>de</strong>s 2 et 3. Les individus en provenance <strong>de</strong> la zone <strong>de</strong> contact 9°50’N, se distribuent<br />

221


Chapitre 4<br />

dans les 3 cla<strong>de</strong>s. La position du cla<strong>de</strong> 3 dans l’arbre phylogénétique montre que ce cla<strong>de</strong><br />

dérive directement du cla<strong>de</strong> 2. Les cla<strong>de</strong>s 1-2, 1-3 et 2-3, présentent respectivement une<br />

divergence <strong>de</strong> 0,5%, 0,9% et 0,2%. La distribution <strong>de</strong>s cla<strong>de</strong>s au sein <strong>de</strong> chaque site (Fig.<br />

4.24.B) révèle que :<br />

(1) les populations <strong>de</strong>s champs 21°N et 13°N sont composées uniquement du<br />

cla<strong>de</strong> 1,<br />

(2) la population du champ 9°50’N compte <strong>de</strong>s séquences <strong>de</strong>s 3 cla<strong>de</strong>s avec<br />

une majorité <strong>de</strong> séquences du cla<strong>de</strong> 2,<br />

(3) les populations <strong>de</strong>s champs situés entre 14°S et 17°34’S présentent un<br />

nombre à peu près égal <strong>de</strong> séquences <strong>de</strong>s cla<strong>de</strong>s 2 et 3<br />

(4) la population du champ 21°33’S est constituée presque exclusivement <strong>de</strong><br />

séquences du cla<strong>de</strong> 3 (une seule séquence recapturée dans le cla<strong>de</strong> 2).<br />

Hormis les individus échantillonnés à 9°50’N, aucun individu ayant ses <strong>de</strong>ux allèles<br />

recapturés n’a été détecté hybri<strong>de</strong> (i.e. 1 allèle dans le cla<strong>de</strong> 1 et le <strong>de</strong>uxième dans le cla<strong>de</strong><br />

2 ou 3). Les individus présentant <strong>de</strong>ux allèles recapturés confirment la distribution<br />

géographique observée sur le jeu <strong>de</strong> qualité 3. Quatre <strong>de</strong>s six individus (9°50’N) présentant<br />

les <strong>de</strong>ux types <strong>de</strong> mitochondrie ont été recapturés. Parmi ces 4 individus, les <strong>de</strong>ux allèles<br />

n’ont été recapturés que <strong>pour</strong> 1 individu et sont distribués dans les cla<strong>de</strong>s 1 et 2,<br />

respectivement. Pour les 3 autres individus n’ayant qu’un allèle recapturé, ces allèles se<br />

distribuent dans le cla<strong>de</strong> 1 <strong>pour</strong> <strong>de</strong>ux individus et dans le cla<strong>de</strong> 2 <strong>pour</strong> le troisième<br />

individu.<br />

222


Chapitre 4<br />

A<br />

B<br />

Fig. 4.24 Arbre phylogénétique <strong>de</strong> qualité 3 (A) et distribution géographique <strong>de</strong>s cla<strong>de</strong>s<br />

(B) du gène Lysozyme.<br />

Sur l’arbre phylogénétique :<br />

En bleu, individus à mitochondrie <strong>de</strong> type ‘espèce Nord’. En rouge, individus à<br />

mitochondrie <strong>de</strong> type ‘espèce Sud’. En noirs, individus à <strong>de</strong>ux types <strong>de</strong> mitochondries.<br />

Sur la distribution <strong>de</strong>s cla<strong>de</strong>s :<br />

Chaque couleur (bleue, rouge, jaune) correspond aux séquences échantillonnées dans les<br />

cla<strong>de</strong>s (1, 2 et 3, respectivement) <strong>de</strong> l’arbre phylogénétique.<br />

223


Chapitre 4<br />

L’arbre phylogénétique <strong>de</strong>s allèles du gène VDG3 (Fig. 4.25.A) révèle également 3<br />

cla<strong>de</strong>s. Les divergences entre cla<strong>de</strong>s 1-2, 1-3 et 2-3 sont respectivement <strong>de</strong> 2,8%, 4% et<br />

1,9%. Le cla<strong>de</strong> 1 compte en majorité <strong>de</strong>s séquences <strong>de</strong> la lignée mitochondriale Nord mais<br />

également quelques allèles appartenant à la lignée mitochondriale Sud ou aux individus<br />

ayant les <strong>de</strong>ux types <strong>de</strong> mitochondrie. Au contraire, les cla<strong>de</strong>s 2 et 3 contiennent<br />

uniquement <strong>de</strong>s séquences recapturées <strong>de</strong> la lignée mitochondriale Sud (mais une séquence<br />

<strong>de</strong> 13°N éliminée lors <strong>de</strong> la sélection aléatoire <strong>de</strong>s séquences présentait une séquence <strong>de</strong><br />

type cla<strong>de</strong> 3 et une mitochondrie <strong>de</strong> type Nord). La distribution <strong>de</strong>s cla<strong>de</strong>s dans les<br />

différentes populations (Fig. 4.25.B) révèlent que, si les 3 cla<strong>de</strong>s sont représentés en<br />

fréquence variable dans la majorité <strong>de</strong>s populations, seules <strong>de</strong>s séquences du cla<strong>de</strong> 1 ont<br />

été échantillonnées à 21°N, 13°N et 17°25’S (hormis la séquence <strong>de</strong> 13°N échantillonnée<br />

dans le cla<strong>de</strong> 3 mais ayant été aléatoirement éliminée dans le jeu <strong>de</strong> données <strong>de</strong> qualité 3).<br />

224


Chapitre 4<br />

A<br />

B<br />

Fig. 4.25 Arbre phylogénétique (A) et distribution géographique <strong>de</strong>s cla<strong>de</strong>s (B) du gène<br />

VDG3.<br />

Sur l’arbre phylogénétique :<br />

En bleu, individus à mitochondrie <strong>de</strong> type ‘espèce Nord’. En rouge, individus à<br />

mitochondrie <strong>de</strong> type ‘espèce Sud’. En noir, individus à <strong>de</strong>ux types <strong>de</strong> mitochondries.<br />

Sur la distribution <strong>de</strong>s cla<strong>de</strong>s ;<br />

Chaque couleur (bleue, rouge, jaune) correspond aux séquences échantillonnées dans les<br />

cla<strong>de</strong>s (1, 2 et 3, respectivement) <strong>de</strong> l’arbre phylogénétique.<br />

225


Chapitre 4<br />

L’arbre phylogénétique <strong>de</strong>s allèles du gène Textilinine peut être divisé en <strong>de</strong>ux<br />

cla<strong>de</strong>s qui divergent <strong>de</strong> 0,6% (Fig. 4.26). Le cla<strong>de</strong> 1 contient l’ensemble <strong>de</strong>s séquences <strong>de</strong>s<br />

individus échantillonnés à 21°N et 13°N, ces individus correspondant à la lignée<br />

mitochondriale nord, plusieurs séquences d’individus <strong>de</strong> 9°50’N (appartenant à la lignée<br />

mitochondriale sud ou ayant les <strong>de</strong>ux types <strong>de</strong> mitochondrie) et une séquence <strong>de</strong> 17°25’S<br />

appartenant à la lignée mitochondriale sud. Le cla<strong>de</strong> 2 présente quant à lui uniquement <strong>de</strong>s<br />

séquences appartenant à la lignée mitochondriale Sud échantillonnées à 9°50’N et 21°33’S<br />

et un allèle issu d’une lignée mitochondriale nord échantillonnée à 9°50N.<br />

Localité N introg N<br />

13°N 0 9<br />

9°50’N 5 10<br />

14°S 0 1<br />

17°25’S 1 3<br />

21°33’S 0 4<br />

Fig. 4.26 Arbre phylogénétique du gène Textilinine et nombre d’individus introgressés<br />

(tableau).<br />

En bleu, séquences d’individus <strong>de</strong> mitochondrie <strong>de</strong> type espèce Nord.<br />

En rouge, séquences d’individus <strong>de</strong> mitochondrie <strong>de</strong> type espèce Sud.<br />

En noir, séquences d’individus présentant <strong>de</strong>ux types <strong>de</strong> mitochondries (Nord et Sud).<br />

N introg : nombre d’individus introgressés, N : nombre total d’individus<br />

226


Chapitre 4<br />

Les valeurs <strong>de</strong> différenciation génétique (φ st ) ont été calculées <strong>pour</strong> chaque gène sur<br />

la totalité <strong>de</strong>s populations puis entre paires <strong>de</strong> populations voisines <strong>pour</strong> les gènes<br />

Lysozyme et VDG3 via le logiciel Arlequin 3.1 (Excoffier et al 2005). Pour le gène<br />

Textilinine, le nombre <strong>de</strong> séquences recapturées <strong>pour</strong> le gène unique étant faible <strong>pour</strong><br />

certaines populations, les φ st ont été calculés en regroupant les populations suivant les<br />

barrières aux flux <strong>de</strong> gènes détectées sur le gène mtCOI (13°N-9°50’N et à l’Equateur),<br />

c’est-à-dire : (1) 13°N (les individus du site 21°N n’ayant pas pu être amplifiés), (2)<br />

9°50’N et (3) 14°S à 21°33’S. La significativité <strong>de</strong> l’écart <strong>de</strong> chaque valeur <strong>de</strong> φ st vis-à-vis<br />

<strong>de</strong> zéro est évaluée par un test <strong>de</strong> 1000-permutations par ce même logiciel.<br />

Les φ st calculés <strong>pour</strong> chaque gène sur l’ensemble <strong>de</strong>s populations sont significatifs<br />

(φ st Lysozyme = 0,518***, φ st VDG3 = 0,235***, φ st Textilinine = 0,267***). Les φ st entre chaque<br />

paire <strong>de</strong> populations voisines (Fig. 4.27) montrent, aux locus Lysozyme et VDG3 , 2 zones<br />

<strong>de</strong> différenciation génétique significative, l’une située entre 21°33’S et 17°34’S, et l’autre<br />

entre 9°50’N et 13°N. Le gène VDG3 présente également un φ st significatif entre les<br />

populations <strong>de</strong> part et d’autre <strong>de</strong> l’Equateur (i.e. 9°50’N et 14°S), mais pas le gène<br />

Lysozyme. Pour le gène Textilinine, les φ st calculés entre 13°N et 9°50’N (φ st = 0,007)<br />

ainsi qu’entre 9°50’N et le Sud <strong>de</strong> l’EPR (φ st = 0,014) ne sont pas significatifs. Cependant,<br />

le φ st entre 13°N et le Sud <strong>de</strong> l’EPR est significatif (p = 0,046*).<br />

227


Chapitre 4<br />

Fig. 4.27 Valeurs <strong>de</strong> φ st entre paires <strong>de</strong> populations voisines sur les gènes Lysozyme<br />

(bleu) et VDG3 (rouge).<br />

* p-value < 0,05 ; ** p < 0,01 ; *** p < 0,001<br />

228


Chapitre 4<br />

Les gènes nucléaires Lysozyme et Textiline présentant une forte ségrégation <strong>de</strong> leurs<br />

allèles selon les <strong>de</strong>ux types mitochondriaux, une étu<strong>de</strong> <strong>de</strong>s mauvaises assignations d’un<br />

type mitochondrial à son ‘background’ nucléaire a été effectuée. Bien évi<strong>de</strong>mment, il<br />

s’agit d’une estimation biaisée par la recapture <strong>de</strong>s allèles, les individus n’ayant pas été<br />

vraiment génotypés aux locus nucléaires (allèles manquants). Un individu est alors<br />

considéré introgressé si au moins un <strong>de</strong> ses allèles nucléaires est échantillonné dans un<br />

cla<strong>de</strong> ne correspondant pas à l’assignation <strong>de</strong> la lignée mitochondriale (e.g. mtCOI Sud et<br />

un allèle dans le cla<strong>de</strong> 1 <strong>pour</strong> le gène Lysozyme et/ou Textilinine). Le gène VDG3 ne<br />

présentant pas une correspondance claire entre les cla<strong>de</strong>s et les lignées mitochondriales, il<br />

n’a pas été utilisé <strong>pour</strong> cette estimation. Le nombre d’individus détectés hybri<strong>de</strong>s ou<br />

introgressés entre les <strong>de</strong>ux espèces cryptiques est faible (7 individus soit 6%, Tableau<br />

4.15). Ces individus proviennent en majorité (6/7) <strong>de</strong> la zone <strong>de</strong> contact 9°50’N (seul 1<br />

Tableau 4.15 Nombre d’individus considérés introgressés, en comparant les cla<strong>de</strong>s<br />

mtCOI aux cla<strong>de</strong>s <strong>de</strong>s gènes Lysozyme et Textilinine, selon leur type mitochondrial et<br />

leur latitu<strong>de</strong>.<br />

Nombre d’individus détectés introgressés/nombre d’individus considérés.<br />

- : absence du type mitochondrial considéré (ou non échantillonné)<br />

individu détecté introgressé à 17°25’S).<br />

Latitu<strong>de</strong> mtCOI Nord mtCOI Sud mtCOI Nord et Sud<br />

20°N 0/11 - -<br />

13°N 0/22 - -<br />

9°50'N 1/2 3/14 2/5<br />

14°S - 0/11 -<br />

17°25'S - 1/10 -<br />

17°34'S - 0/12 -<br />

21°33'S - 0/19 -<br />

229


Chapitre 4<br />

4.3.3.3. Rôle <strong>de</strong>s barrières aux flux <strong>de</strong> gènes dans la dynamique <strong>de</strong><br />

l’introgression<br />

Le contact secondaire entre <strong>de</strong>ux lignées divergentes géographiquement séparées<br />

peut, lorsque la barrière reproductive n’est pas totale, donner lieu à <strong>de</strong> l’hybridation entre<br />

ces lignées. Si la dispersion est suffisamment efficace et que les hybri<strong>de</strong>s ne sont pas<br />

contre-sélectionnés, cette remise en contact secondaire tendra vers une homogénéisation<br />

génétique rapi<strong>de</strong> <strong>de</strong> ces lignées. En revanche, un équilibre entre dispersion et sélection<br />

contre les hybri<strong>de</strong>s peut maintenir une zone d’hybridation <strong>de</strong> largeur variable (Barton &<br />

Hewitt 1985). Ainsi, alors que la zone hybri<strong>de</strong> entre les moules côtières Mytilus edulis et<br />

M. galloprovincialis se distribue en mosaïque sur <strong>de</strong>s centaines <strong>de</strong> kilomètres en formant<br />

une succession <strong>de</strong> zones <strong>de</strong> transition entre le Nord et le Sud <strong>de</strong> l’Atlantique (Bierne et al<br />

2003), la zone « hybri<strong>de</strong> » <strong>de</strong> la sauge buissonnante Artemisia tri<strong>de</strong>ntata se limite à un<br />

front d’une quinzaine <strong>de</strong> mètres entre les aires géographiques <strong>de</strong>s 2 lignées (Graham et al<br />

1995). La détection d’individus hybri<strong>de</strong>s entre les <strong>de</strong>ux espèces cryptiques <strong>de</strong> L. elevatus<br />

définies par Matabos et al (2008b) à 9°50’N (via <strong>de</strong>s marqueurs enzymatiques), associée à<br />

la distribution géographique <strong>de</strong> ces espèces (précisée par Johnson et al (2008) et Plouviez<br />

et al (2009) sur le mtCOI), suggérait que le champ hydrothermal 9°50’N pouvait constituer<br />

une zone <strong>de</strong> tension dont l’étendue, le mo<strong>de</strong> <strong>de</strong> fonctionnement, et l’impact sur l’isolement<br />

génétique <strong>de</strong>s espèces restaient à établir.<br />

La distribution géographique (21°N-9°50’N et 9°50’N-21°S, respectivement) <strong>de</strong>s<br />

allèles appartenant aux <strong>de</strong>ux cla<strong>de</strong>s divergents sur chacun <strong>de</strong>s gènes Lysozyme et<br />

Textilinine concor<strong>de</strong> avec la distribution <strong>de</strong>s cla<strong>de</strong>s mtCOI <strong>de</strong>s <strong>de</strong>ux espèces et confirme la<br />

présence d’une zone <strong>de</strong> contact secondaire à 9°50’N. De plus, la majorité <strong>de</strong>s individus<br />

possédant une signature ‘hybri<strong>de</strong>’ avec un allèle ‘nord’ et un allèle ‘sud’ découlant <strong>de</strong><br />

chacun <strong>de</strong>s cla<strong>de</strong>s (hybri<strong>de</strong>s ou introgressés) a été échantillonnée sur ce champ<br />

hydrothermal. Le gène VDG3 présente également <strong>de</strong>s cla<strong>de</strong>s divergents. Cependant, la<br />

distribution géographique <strong>de</strong> ces cla<strong>de</strong>s reste difficile à établir et ne correspond pas aux<br />

patrons <strong>de</strong> distribution observés aux 3 autres gènes. En effet, si les cla<strong>de</strong>s 2 et 3 ne<br />

comptent que <strong>de</strong>s séquences échantillonnées au sein <strong>de</strong>s individus <strong>de</strong> l’espèce sud, le cla<strong>de</strong><br />

1 correspond à un mélange équiprobable <strong>de</strong> séquences échantillonnées dans les 2 lignées<br />

230


Chapitre 4<br />

mitochondriales trouvées au Nord et au Sud <strong>de</strong> l’EPR. De ce fait, il est difficile <strong>de</strong> savoir si<br />

(1) le cla<strong>de</strong> 1 correspond à la lignée Nord, une partie <strong>de</strong>s allèles ‘nord’ ayant introgressé la<br />

lignée sud dans un passé relativement ancien (certains allèles ‘introgressés’ ayant eu le<br />

temps <strong>de</strong> se diversifier dans la lignée sud) ou (2) si cette distribution <strong>de</strong>s allèles n’est pas<br />

uniquement le reflet d’un polymorphisme ‘ancestral’ non trié (i.e. séparation <strong>de</strong>s lignées<br />

n’ayant pas permis <strong>de</strong> fixer suffisamment <strong>de</strong> mutations dans la divergence <strong>pour</strong> atteindre<br />

une monophylie réciproque sur ce gène). La distinction entre une introgression ancienne<br />

d’allèles et un polymorphisme ancestral n’est pas chose aisée, lorsque l’arbre<br />

phylogénétique présente une para- ou une polyphylie (Nosil et al 2009), d’autant plus que<br />

les gènes nucléaires nécessitent un nombre plus important <strong>de</strong> générations <strong>pour</strong> atteindre<br />

cette monophylie réciproque (Zink & Barrowclough 2008), en comparaison <strong>de</strong>s locus<br />

mitochondriaux. En tout état <strong>de</strong> cause, il semble néanmoins que, si le mélange d’allèles<br />

‘sud’ et ‘nord’ dans le cla<strong>de</strong> 1 reflète une introgression, alors celle-ci doit être relativement<br />

ancienne, car certaines branches terminales échantillonnées dans la lignée sud coalescent<br />

relativement profondément dans le cla<strong>de</strong> ‘nord’ <strong>de</strong> l’arbre phylogénétique, indiquant une<br />

accumulation <strong>de</strong> polymorphisme <strong>de</strong>puis l’événement d’introgression potentiel (Fig.<br />

4.25.A). Dans ce cas, l’ancienneté <strong>de</strong> l’événement <strong>pour</strong>rait expliquer <strong>pour</strong> ce gène,<br />

l’importance <strong>de</strong> la zone géographique balayée par l’introgression au sud <strong>de</strong> l’EPR (allèles<br />

‘introgressés’ retrouvés jusqu’à 21°33S), si tant est que l’on considère le gène VDG3<br />

comme étant un marqueur neutre. Une telle hypothèse apparaît relativement réaliste même<br />

<strong>pour</strong> une espèce à faible capacité dispersive si l’on considère que l’événement <strong>de</strong><br />

spéciation entre les lignées sud et nord date d’au moins 10 millions d’années (Plouviez et<br />

al 2009) et que la remise en contact entre les 2 lignées <strong>pour</strong>raient largement pré-dater la<br />

mise en place <strong>de</strong> la secon<strong>de</strong> barrière équatoriale, il y a 1,3 millions d’années. À l’inverse,<br />

le faible nombre d’individus ayant une signature ‘hybri<strong>de</strong>’ sur les gènes Lysozyme et<br />

Textilinine et leur position géographique relativement restreinte (essentiellement à 9°50’N)<br />

<strong>pour</strong>raient être la conséquence d’une hybridation récente <strong>de</strong>s 2 lignées (durée<br />

d’introgression trop courte <strong>pour</strong> être détectée au niveau <strong>de</strong>s cla<strong>de</strong>s) ou la conséquence<br />

d’une trop forte contre-sélection <strong>de</strong>s individus hybri<strong>de</strong>s (barrière génétique efficace <strong>de</strong> part<br />

et d’autre <strong>de</strong> 9°50N) qui masquerait, à ces locus, l’hybridation ancienne <strong>de</strong> ces 2 lignées. Il<br />

convient cependant <strong>de</strong> noter que la plupart <strong>de</strong> ces individus ne sont pas hybri<strong>de</strong>s <strong>de</strong><br />

première génération (introgression), indiquant que la contre-sélection <strong>de</strong>s hybri<strong>de</strong>s n’est<br />

pas totale.<br />

231


Chapitre 4<br />

Si la sélection contre les hybri<strong>de</strong>s peut avoir jouer un rôle important dans la<br />

limitation géographique <strong>de</strong> l’hybridation <strong>pour</strong> certains gènes, la barrière à la dispersion<br />

récemment mise en place entre 9°50’N et 14°S (1,3 Ma, Plouviez et al 2009), mise en<br />

évi<strong>de</strong>nce au locus mitochondrial et confirmée sur le gène Lysozyme, <strong>pour</strong>rait avoir<br />

récemment réduit le flux migratoire <strong>de</strong> l’espèce Sud entre 9°50’N et le Sud EPR, limitant<br />

par la même, le contact secondaire entre les espèces à la zone située à 9°50’N et donc,<br />

l’étendue <strong>de</strong> la zone d’introgression aux nouveaux allèles apparus <strong>de</strong>puis la mise en place<br />

<strong>de</strong> la secon<strong>de</strong> barrière (la zone d’introgression beaucoup plus étendue observée au gène<br />

VDG3 étant probablement le reflet d’une histoire plus ancienne que la mise en place <strong>de</strong><br />

cette barrière). Ainsi, la population <strong>de</strong> la lignée sud échantillonnée sur le champ<br />

hydrothermal 9°50’N serait partiellement isolée <strong>de</strong> celles <strong>de</strong>s autres champs<br />

hydrothermaux situés plus au Sud, pouvant ainsi expliquer l’étroitesse <strong>de</strong> la zone<br />

d’hybridation vers le Nord et le Sud <strong>pour</strong> les gènes Lysozyme et Textilinine malgré une<br />

histoire évolutive plus ancienne <strong>de</strong>s 2 lignées.<br />

La mise en place récente (1,3 Ma) <strong>de</strong> la barrière à l’équateur <strong>pour</strong>rait également être à<br />

l’origine <strong>de</strong> la présence <strong>de</strong>s 2 cla<strong>de</strong>s divergents retrouvés uniquement au Sud <strong>de</strong> l’EPR (i.e.<br />

cla<strong>de</strong>s 2 et 3) <strong>pour</strong> les gènes Lysozyme et VDG3. En effet, les cla<strong>de</strong>s 2 <strong>de</strong>s 2 gènes<br />

<strong>pour</strong>raient correspondre à l’isolement <strong>de</strong> la population 9°50’N au Nord <strong>de</strong> la barrière alors<br />

que les cla<strong>de</strong>s 3 correspondraient aux populations isolées au sud <strong>de</strong> la barrière <strong>de</strong> 14°S à<br />

21°33’S. Cette hypothèse s’appuie sur le fait que la divergence entre les cla<strong>de</strong>s 1 et 2 est<br />

plus faible que celle trouvée entre les cla<strong>de</strong>s 1 et 3 <strong>pour</strong> les 2 gènes, et que le cla<strong>de</strong> 3 est<br />

majoritairement retrouvé au sud <strong>de</strong> la barrière 9°50’N-14°S. Une remise en contact <strong>de</strong>s<br />

populations <strong>de</strong> ces cla<strong>de</strong>s 2 et 3 auraient ensuite pu brouiller le signal d’isolement<br />

géographique entre ces populations (pas <strong>de</strong> monophylie réciproque) laissant supposer que<br />

cette secon<strong>de</strong> barrière n’est pas très étanche <strong>pour</strong> ce gastéropo<strong>de</strong> hydrothermal.<br />

232


Chapitre 4<br />

4.3.3.4. Isolement génétique <strong>de</strong>s populations <strong>de</strong> 21°N et 21°33’S et<br />

sens <strong>de</strong> colonisation<br />

Outre l’isolement génétique <strong>de</strong>s populations <strong>de</strong> part et d’autre <strong>de</strong>s barrières génétique<br />

13°N-9°50’N et physique 9°50’N-14°S observées, les populations situées en limite d’aire<br />

géographique <strong>de</strong> l’espèce (21°N et 21°33S) présentent un particularisme génétique sur au<br />

moins l’un <strong>de</strong>s gènes échantillonnés : la population 21°N se distingue par une réduction <strong>de</strong><br />

polymorphisme au locus mtCOI (différence significative avec la population 13°N,<br />

Plouviez et al 2009) et la population 21°33’S présente une série d’allèles relativement<br />

divergents en forte fréquence aux gènes nucléaires Lysozyme et Textilinine (différence<br />

significative avec 17°34’S). Outre un décalage géographique <strong>de</strong>s isolats génétiques<br />

associés à la mise en place <strong>de</strong> la secon<strong>de</strong> barrière, ces observations <strong>pour</strong>raient également<br />

traduire <strong>de</strong>s événements démographiques locaux.<br />

Ne comptant qu’un unique haplotype mitochondrial, indiquant une réduction<br />

drastique <strong>de</strong> la diversité génétique <strong>de</strong> la population 21°N, la différenciation génétique entre<br />

13°N et 21°N <strong>pour</strong>rait être due à un goulot d’étranglement <strong>de</strong> la population <strong>de</strong> 21°N suite à<br />

une extinction massive <strong>de</strong>s émissions qui aurait pu avoir lieu au début <strong>de</strong>s années 80<br />

(Hessler et al 1985). Ce résultat suggère ainsi que la recolonisation <strong>de</strong>s nouvelles zones<br />

d’émission ne s’effectuerait que très partiellement à partir <strong>de</strong>s sites situées plus au sud sur<br />

l’EPR, très probablement en raison <strong>de</strong> la présence <strong>de</strong> la faille transformante Rivera qui est<br />

l’une <strong>de</strong>s plus longues du système EPR (décalage <strong>de</strong> la dorsale sur plus <strong>de</strong> 500 kilomètres).<br />

Il convient néanmoins <strong>de</strong> noter que Shank & Halanych (2007) ont suggéré que <strong>de</strong>s<br />

immigrations récentes <strong>de</strong> Riftia pachyptila en provenance <strong>de</strong> 9°50N étaient détectables<br />

dans les populations du bassin <strong>de</strong> Guaymas par utilisation <strong>de</strong> marqueurs AFLP. Cependant,<br />

aucune réduction <strong>de</strong> diversité n’a pu être notée sur les gènes Lysozyme et VDG3 au niveau<br />

<strong>de</strong> la population 21°N, comme on l’attendrait <strong>pour</strong> un événement démographique.<br />

Aucune faille transformante majeure n’est présente entre 17°34’S et 21°33’S,<br />

indiquant que la différenciation entre ces populations ne serait pas liée à l’existence d’une<br />

barrière tectonique surnuméraire. La dynamique <strong>de</strong>s courants profonds (<strong>de</strong> l’ouest vers<br />

l’est à ces latitu<strong>de</strong>s et quasi-perpendiculaire à l’axe du graben, Reid 1997) <strong>pour</strong>rait être<br />

responsable <strong>de</strong> l’isolement <strong>de</strong> ces populations. Cependant, une telle différenciation entre la<br />

population <strong>de</strong> 21°33S et les autres populations <strong>de</strong> l’EPR sud n’a actuellement pas été mise<br />

233


Chapitre 4<br />

en évi<strong>de</strong>nce chez les autres espèces. Si une migration du Sud vers le Nord à travers la<br />

barrière équatoriale a été détectée sur la lignée Sud <strong>de</strong> l’espèce L. elevatus <strong>pour</strong> le gène<br />

COI (Plouviez et al 2009), l’arbre phylogénétique du gène Lysozyme suggère que la<br />

population 21°33’S <strong>pour</strong>rait être dérivée d’un effet <strong>de</strong> fondation à partir <strong>de</strong>s populations<br />

situées plus au Nord (e.g. Animal Farm : 18°33S). En effet, le cla<strong>de</strong> 3 <strong>de</strong> ce gène,<br />

majoritaire à 21°33’S, est dérivé du cla<strong>de</strong> 2 échantillonné en forte fréquence sur les autres<br />

populations <strong>de</strong> la lignée sud (9°50’N à 17°34’S).<br />

Avec la détection <strong>de</strong>s <strong>de</strong>ux barrières aux flux <strong>de</strong> gènes (13°N-9°50’N et 9°50’N-<br />

14°S) et la différenciation génétique <strong>de</strong>s populations 21°N et 21°33’S, L. elevatus semble<br />

présenter une structure génétique plus marquée que les 2 autres espèces étudiées dans ce<br />

chapitre (une seule barrière détectée au sein <strong>de</strong>s espèces A. pompejana et B. thermophilus).<br />

Cette structure génétique plus forte <strong>pour</strong>rait indiquer que l’espèce a <strong>de</strong>s capacités <strong>de</strong><br />

dispersion plus restreintes et/ou une sensibilité plus forte aux modifications <strong>de</strong><br />

l’environnement local (extinction). Il est néanmoins important <strong>de</strong> considérer ici que<br />

l’histoire évolutive plus complexe <strong>de</strong> L. elevatus (2 barrières et 2 pério<strong>de</strong>s d’isolement)<br />

peut également fournir une explication convaincante à cette structure génétique plus<br />

marquée. L’hypothèse d’une dispersion relativement limitée (type proche en proche) <strong>de</strong><br />

l’espèce fournit cependant un élément <strong>de</strong> compréhension supplémentaire <strong>pour</strong> expliquer le<br />

nombre relativement important d’espèces décrites chez les Lepetodrilidae (famille pouvant<br />

franchir plus difficilement les barrières à la dispersion et ainsi diverger plus rapi<strong>de</strong>ment <strong>de</strong><br />

part et d’autre d’une barrière par spéciation allopatrique). En effet, la diversification<br />

d’espèces en allopatrie partiellement due aux capacités <strong>de</strong> dispersion réduites <strong>de</strong>s larves<br />

(non-planctotrophes), a été mise en évi<strong>de</strong>nce chez <strong>de</strong>s gastéropo<strong>de</strong>s marins du genre Conus<br />

(Cuhna et al 2005). La radiation <strong>de</strong>s Lepetodrilidae <strong>pour</strong>rait ainsi être le reflet d’une<br />

dispersion larvaire limitée. Des étu<strong>de</strong>s plus approfondies sur les capacités <strong>de</strong> dispersion <strong>de</strong><br />

ces espèces <strong>pour</strong>raient ainsi compléter ces informations génétiques et permettre <strong>de</strong> mieux<br />

comprendre la radiation <strong>de</strong> cette famille. Néanmoins, il est fort probable qu’une part <strong>de</strong><br />

cette radiation évolutive soit également issue <strong>de</strong> spéciations adaptatives (spécialisation<br />

écologique).<br />

234


Chapitre 4<br />

4.3.4. Y a-t-il eu une expansion démographique au Sud <strong>de</strong> l’EPR ?<br />

4.3.4.1. Indices <strong>de</strong> diversité et tests d’écart à la neutralité<br />

La diversité nucléotidique (π n ) et le thêta <strong>de</strong> Watterson (θ W ) ont été calculés via le<br />

logiciel DNAsp 4.10.3 (Rozas et al 2003) au sein <strong>de</strong> chaque cla<strong>de</strong> <strong>pour</strong> chaque gène et sont<br />

présentés dans le tableau 4.16. Pour ces gènes, si la recombinaison in vitro détectée entre<br />

individus a pu être éliminée, et que le nombre <strong>de</strong> recombinants inter-allèles au sein du<br />

même individu ait été minimisé par suppression <strong>de</strong>s allèles surnuméraires, la<br />

recombinaison naturelle non détectée peut néanmoins persister dans nos jeux <strong>de</strong> séquences<br />

et est souvent visible lorsque l’on regar<strong>de</strong> la longueur <strong>de</strong>s branches terminales <strong>de</strong> certains<br />

allèles. Le ‘four-gamete test’ <strong>de</strong> Hudson & Kaplan (1985), visant à détecter le nombre<br />

minimum d’événements <strong>de</strong> recombinaison (R m ), a donc été effectué <strong>pour</strong> chacun <strong>de</strong>s gènes.<br />

Les R m calculées <strong>pour</strong> chaque gène sur la totalité <strong>de</strong>s séquences montrent <strong>de</strong>s valeurs non<br />

nulles, allant <strong>de</strong> 5 <strong>pour</strong> le gène VDG3 à 9 <strong>pour</strong> les <strong>de</strong>ux autres gènes nucléaires, indiquant<br />

la persistance d’allèles recombinés dans nos jeux <strong>de</strong> données. Néanmoins, il n’était pas<br />

possible <strong>de</strong> supprimer ces séquences potentiellement recombinantes afin d’obtenir un R m<br />

global nul car cela impliquait une suppression d’un nombre trop important <strong>de</strong> séquences.<br />

Les logiciels basés sur la coalescence faisant l’hypothèse <strong>de</strong> l’absence <strong>de</strong> recombinaison<br />

(e.g. Migrate-n, IM, Sweep_bott), ils n’ont donc pas été utilisés dans cette étu<strong>de</strong>. Afin <strong>de</strong><br />

détecter d’éventuels écarts à la neutralité, les indices <strong>de</strong> Tajima (D, 1989) et <strong>de</strong> Fu & Li<br />

(F*, 1993) ont été réalisés <strong>pour</strong> chacun <strong>de</strong>s cla<strong>de</strong>s, leur valeur moyenne n’étant<br />

théoriquement pas affectée par la recombinaison (seule la variance est impactée, Schierup<br />

& Hein 2000). Cependant, la recombinaison, en augmentant le nombre <strong>de</strong> mutations en<br />

fréquence intermédiaire, influence la distribution <strong>de</strong>s différences par paire <strong>de</strong> séquences<br />

(courbes <strong>de</strong> mismatch, Fig. 4.28).<br />

235


Chapitre 4<br />

Tableau 4.16 Diversité nucléotidique (π n ), thêta <strong>de</strong> Watterson (θ W ), indices <strong>de</strong> Tajima<br />

(D) et <strong>de</strong> Fu & Li (F*), et nombre minimum d’événement <strong>de</strong> recombinaison (R m ) au<br />

sein <strong>de</strong> chaque cla<strong>de</strong> <strong>pour</strong> les gènes nucléaires.<br />

N, nombre <strong>de</strong> séquences (correspondant également au nombre d’individus)<br />

NS p > 0,10 ; + p < 0,10 ; * p < 0,05 ; ** p < 0,01 ; *** p < 0,001<br />

Locus Cla<strong>de</strong> Ν π n θ W D F* R m<br />

Lysozyme 1 33 0,011 0,017 -1,387 NS -2,263 + 8<br />

2 28 0,001 0,005 -2,445 ** -4,263 ** 0<br />

3 40 0,002 0,008 -2,554 *** -4,644 ** 0<br />

VDG3 1 68 0,005 0,010 -1,675 + -3,324 ** 1<br />

2 14 0,003 0,004 -1,573 NS -2,135 + 1<br />

3 8 0,002 0,002 -1,064 NS -1,336 NS 0<br />

Textilinine 1 17 0,014 0,020 -1,305 NS -1,747 NS 4<br />

2 13 0,008 0,012 -1,420 NS -2,197 + 2<br />

236


Chapitre 4<br />

Fig. 4.28 Histogrammes <strong>de</strong> fréquence du nombre <strong>de</strong> différences par paire <strong>de</strong> séquences<br />

au sein <strong>de</strong> chaque cla<strong>de</strong> <strong>pour</strong> les 3 gènes nucléaires.<br />

Traits pointillés, courbes observées. Traits pleins, courbes attendues sous l’hypothèse d’une<br />

homogénéité du cla<strong>de</strong> considéré et <strong>de</strong> la neutralité.<br />

237


Chapitre 4<br />

4.3.4.2. Expansion démographique au Sud<br />

Matabos et al (2008b) ont suggéré l’existence d’un goulot d’étranglement récent<br />

dans la lignée sud <strong>de</strong> L. elevatus à 17°25’S en comparant, sur <strong>de</strong>s marqueurs allozymiques<br />

et <strong>de</strong>s marqueurs DALP, l’hétérozygotie attendue observée à celle attendue sous<br />

l’hypothèse mutation/dérive (software Bottleneck, Cornuet & Luitkart 1996). Une<br />

expansion démographique plus globale a été ensuite mise en évi<strong>de</strong>nce sur 7 espèces le long<br />

<strong>de</strong> l’EPR Sud, sur le gène mtCOI, et datée à moins <strong>de</strong> 500 000 ans (Plouviez et al 2009),<br />

les valeurs <strong>de</strong> D et F* obtenues chez la lignée sud <strong>de</strong> L. elevatus étant en accord avec cette<br />

hypothèse. Les gènes nucléaires utilisés sur les espèces A. pompejana et B. thermophilus<br />

(cf. parties 1 et 2 <strong>de</strong> ce chapitre) ont confirmé ce résultat. Chez L. elevatus, les valeurs<br />

significativement négatives du D <strong>de</strong> Tajima et du F* <strong>de</strong> Fu & Li recensées uniquement<br />

dans la lignée sud ainsi que les courbes <strong>de</strong> mésappariement ‘mismatch’, qui présentent un<br />

excès <strong>de</strong> variants rares (avec un Rm = 0) au Sud sur le gène Lysozyme, sont également en<br />

accord avec cette hypothèse. Pour le gène VDG3, la distribution géographique <strong>de</strong>s cla<strong>de</strong>s<br />

ne permet pas <strong>de</strong> relier les valeurs <strong>de</strong> ces indices au Sud <strong>de</strong> la dorsale. Le gène Textilinine,<br />

qui discrimine bien le Nord et le Sud <strong>de</strong> l’EPR par 2 cla<strong>de</strong>s divergents, présente <strong>de</strong>s indices<br />

<strong>de</strong> Tajima et Fu & Li qui ne sont pas significatifs au Nord et en limite <strong>de</strong> significativité au<br />

Sud (significatifs à 10%). Un échantillonnage plus conséquent <strong>de</strong> ces cla<strong>de</strong>s <strong>pour</strong>rait<br />

permettre <strong>de</strong> mieux évaluer la significativité <strong>de</strong> ces tests. Si la valeur moyenne <strong>de</strong> ces<br />

indices n’est théoriquement pas affectée par la recombinaison (Schierup & Hein 2000), les<br />

courbes <strong>de</strong> mésappariement révèlent quant à elles plusieurs pics <strong>de</strong> variants en fréquence<br />

intermédiaires, reflétant probablement un fort effet <strong>de</strong> la recombinaison <strong>pour</strong> le gène<br />

Textilinine mais beaucoup moins marqués <strong>pour</strong> les gènes VDG3 et Lyso <strong>pour</strong> lesquels un<br />

effet sélectif ou démographique est plus probable. En conséquence, même si les gènes<br />

VDG3 et Textilinine ne permettent pas d’inférer réellement un changement<br />

démographique au sud <strong>de</strong> l’EPR, les autres locus (mtCOI, Lysozyme et allozymes)<br />

s’accor<strong>de</strong>nt en faveur d’une expansion démographique au Sud <strong>de</strong> la dorsale, pouvant être<br />

due à <strong>de</strong> multiple extinctions/recolonisations dans cette région soumise à <strong>de</strong>s vitesses<br />

d’accrétion élevées (125-152 mm/an, Cormier 1997).<br />

238


Chapitre 4<br />

5. Discussion générale du chapitre 4 : comparaison <strong>de</strong>s résultats obtenus<br />

entre les espèces cibles<br />

L’approche multi-locus utilisée dans ce chapitre sur trois espèces (A. pompejana, B.<br />

thermophilus et L. elevatus) appartenant à <strong>de</strong>s classes distinctes d’organismes<br />

Lophotrochozoaires, permet <strong>de</strong> dégager un certain nombre <strong>de</strong> points communs quant à leur<br />

histoire phylogéographique et le rôle <strong>de</strong> barrières physiques à la dispersion le long <strong>de</strong><br />

l’EPR. Quatre points communs majeurs peuvent ainsi être soulignés :<br />

(1) La présence d’un patron commun <strong>de</strong> vicariance entre ces espèces,<br />

géographiquement ciblé entre le nord et le sud <strong>de</strong> la dorsale. Alors que<br />

l’approche multispécifique sur le seul gène mtCOI ne permettait pas <strong>de</strong> définir<br />

précisément la position géographique <strong>de</strong> la barrière aux flux <strong>de</strong> gènes<br />

responsable <strong>de</strong> cette vicariance, l’approche multi-locus a permis <strong>de</strong> préciser la<br />

localisation <strong>de</strong> cette barrière chez B. thermophilus et <strong>de</strong> confirmer la position<br />

préalablement établie sur le mtCOI <strong>pour</strong> les <strong>de</strong>ux autres espèces. De plus, la<br />

datation multi-locus estimée chez A. pompejana et B. thermophilus confirme<br />

celle évaluée via le logiciel MsBayes par l’approche multispécifique à 1,3 Ma.<br />

(2) Le caractère semi-perméable <strong>de</strong> cette barrière aux flux <strong>de</strong> gènes même si celle<br />

ci apparaît relativement étanche <strong>pour</strong> les 3 espèces<br />

(3) L’introgression d’allèles entre les lignées mitochondriales divergentes<br />

(4) La confirmation via l’approche multilocus <strong>de</strong> l’hypothèse d’expansion<br />

démographique au Sud <strong>de</strong> l’EPR proposée au préalable par l’approche<br />

multispécifique. Ainsi, l’effet d’un goulôt d’étranglement suivi d’une forte<br />

expansion <strong>de</strong>s populations observé sur l’EPR Sud semble relativement clair et<br />

lié à un facteur extrinsèque fortement structurant à l’échelle <strong>de</strong> la communauté<br />

et non <strong>de</strong> l’espèce (effet tectonique ou volcanique majeur). Pourtant, révéler un<br />

changement démographique passé à partir d’un déséquilibre mutation/dérive<br />

dans le polymorphisme d’une espèce n’est pas chose aisée et a donné lieu à <strong>de</strong><br />

véritables débats entre partisans <strong>de</strong> l’existence d’un goulot d’étranglement (e.g.<br />

Przeworski et al 2001, Wall et al 2002), ou <strong>de</strong> multiple balayages sélectifs (e.g.<br />

Harr et al 2002, Glinka et al 2003, Baines et al 2002) dans les populations<br />

africaines et non-africaines <strong>de</strong> Drosophila melanogaster. En compilant les locus<br />

239


Chapitre 4<br />

<strong>de</strong> l’ensemble <strong>de</strong>s étu<strong>de</strong>s précé<strong>de</strong>ntes (soient 105 locus) et en y ajoutant 10<br />

locus non codants liés au chromosome X, Haddrill et al (2005) ont alors conclu<br />

à l’existence d’un goulot d’étranglement (certains locus étant également soumis<br />

à diverses formes <strong>de</strong> sélection). Comparer plusieurs locus et plusieurs espèces<br />

co-distribuées peut donc s’avérer extrêmement efficace (et peut être même plus<br />

efficace que d’utiliser <strong>de</strong>s dizaines <strong>de</strong> locus sur une seule espèce) <strong>pour</strong> détecter<br />

<strong>de</strong>s évènements démographiques passés, si ceux-ci sont le fruit d’événements<br />

écologiques ou climatiques majeurs affectant la majorité <strong>de</strong>s espèces d’une<br />

communauté (explosion d’un volcan, changements rapi<strong>de</strong>s du climat<br />

(glaciations), catastrophes <strong>de</strong> type tempête/ouragan ou chute <strong>de</strong> météores).<br />

Mais outre ces points communs entre espèces, l’approche multi-locus a également<br />

permis <strong>de</strong> détecter certaines caractéristiques propres aux traits d’histoire <strong>de</strong> vie <strong>de</strong> chaque<br />

espèce et ainsi <strong>de</strong> révéler <strong>de</strong>s dissimilitu<strong>de</strong>s entre espèces notamment en termes <strong>de</strong><br />

positionnement <strong>de</strong> la barrière et/ou <strong>de</strong> la taille <strong>de</strong>s zones <strong>de</strong> contact. Ces variations, dues à<br />

l’histoire évolutive particulière <strong>de</strong> chaque taxon, notamment au niveau <strong>de</strong> leurs stratégies<br />

<strong>de</strong> reproduction et <strong>de</strong> dispersion, constituent un fort bruitage au signal <strong>de</strong> séparation <strong>de</strong>s<br />

faunes que constitue la barrière. Ainsi, L. elevatus présente une structuration génétique<br />

plus marquée que les <strong>de</strong>ux autres espèces (<strong>pour</strong> lesquelles une seule barrière est détectée),<br />

avec <strong>de</strong>ux événements d’isolement, l’un génétique (création d’une zone <strong>de</strong> tension à 9°50N<br />

suite à une séparation très ancienne <strong>de</strong>s faunes : 11,6 Ma) où le mécanisme principal est la<br />

contre-sélection <strong>de</strong>s hybri<strong>de</strong>s et l’autre physique et plus récente (création d’une barrière<br />

physique à la dispersion entre 9°50’N et 14°S, il y a 1,3 Ma). Ces barrières perméables aux<br />

flux <strong>de</strong> gènes contribuent à une différenciation génétique complexe à laquelle <strong>pour</strong>raient<br />

s’ajouter <strong>de</strong>s effets <strong>de</strong> différenciation localisés au niveau <strong>de</strong>s sites 21°N (extinction locale<br />

du site, il y a quelques dizaines d’années) et 21°33S (point <strong>de</strong> rencontre avec une troisième<br />

espèce cryptique <strong>de</strong> L. elevatus). Cette structuration génétique plus forte <strong>pour</strong>rait<br />

s’expliquer par (1) <strong>de</strong>s capacités <strong>de</strong> dispersion plus faible chez cette espèce et/ou (2) une<br />

plus forte sensibilité à l’extinction <strong>de</strong>s émissions hydrothermales en étant plus<br />

spécifiquement associé au vestimentifère Riftia pachyptila.<br />

Si L. elevatus présente, tout comme les <strong>de</strong>ux autres espèces, une phase <strong>de</strong> dispersion<br />

essentiellement larvaire, les larves sont incubées dans la cavité palléale <strong>de</strong> la femelle<br />

jusqu’au stage veligère avant d’être émises dans la plume hydrothermale (Fretter et al<br />

1988, Mullineaux et al 1995). Les larves <strong>de</strong> cette espèce sont lécitotrophes et n’ont pas été<br />

240


Chapitre 4<br />

décrites dans <strong>de</strong>s couches d’eau éclairées à fortes concentrations en plancton<br />

(contrairement à Bathymodiolus, Arellano & Young 2009). La survie <strong>de</strong>s larves dans une<br />

colonne d’eau oligotrophe telle que celle trouvée au niveau <strong>de</strong> la vallée axiale <strong>de</strong>s dorsales<br />

océaniques implique nécessairement d’importantes réserves vitellines. Or, les œufs <strong>de</strong> L.<br />

elevatus présentent une taille inférieure à ceux d’A. pompejana (qui possè<strong>de</strong> également une<br />

larve lécitotrophe ne migrant, a priori, pas dans la colonne d’eau : Tyler et al 2008). Du<br />

fait d’une incubation <strong>de</strong>s œufs dans la cavité palléale <strong>de</strong>s femelles, d’une part, et <strong>de</strong> la<br />

faible importance <strong>de</strong>s réserves vitellines, d’autre part, on peut donc suggérer que la durée<br />

<strong>de</strong> dispersion dans la colonne d’eau <strong>de</strong>s larves <strong>de</strong> L. elevatus doit être plus faible que celle<br />

supposée chez B. thermophilus ou A. pompejana.<br />

L’hypothèse d’une plus forte sensibilité à l’extinction locale est plus difficile à<br />

appréhen<strong>de</strong>r car peu <strong>de</strong> données sont disponibles quant aux conditions <strong>de</strong> survie <strong>de</strong> cette<br />

espèce. L’espèce L. elevatus présente un preferendum thermique d’environ 4,5-11°C (Mills<br />

et al 2007) et est essentiellement associée aux B. thermophilus et Riftia pachyptila.<br />

Lorsqu’une source perd <strong>de</strong> son activité et se tarît, la température du milieu diminue jusqu’à<br />

revenir à la température <strong>de</strong> l’eau <strong>de</strong> fond (1-5°C), les vestimentifères associés au flui<strong>de</strong><br />

meurent très rapi<strong>de</strong>ment (quelques dizaines <strong>de</strong> jours) alors que les moulières peuvent rester<br />

en place quelques années supplémentaires (évolution du site Parigo 84-90, Jollivet 1993)<br />

même si les modioles per<strong>de</strong>nt rapi<strong>de</strong>ment leurs symbiontes, leurs réserves en glycogène et<br />

leur capacité <strong>de</strong> reproduction (Fiala-Médioni et al 1986, Raulfs et al 2004). Ces<br />

changements faunistiques peuvent ainsi modifier la distribution locale <strong>de</strong> L. elevatus. Pour<br />

atteindre les conditions 'idéales' <strong>de</strong> leur preferendum, les individus <strong>de</strong>vraient donc se<br />

déplacer au plus proche du flui<strong>de</strong>. L. elevatus <strong>pour</strong>rait entrer en compétition avec d’autres<br />

espèces, et notamment L. ovalis, qui est associé plus spécifiquement aux moules<br />

hydrothermales, avec un preferendum thermique plus faible (environ 1,8-6°C, Mills et al<br />

2007). Une telle compétition entre L. elevatus et L. ovalis sur la gamme <strong>de</strong> température<br />

préférentielle commune, concor<strong>de</strong> avec les observations <strong>de</strong> distribution <strong>de</strong> ces espèces. En<br />

effet, alors que L. ovalis a été échantillonné en grand nombre à 7°25’S et 18°33’S (sites en<br />

voie d’extinction avec populations nécrophages), l’espèce L. elevatus y est relativement<br />

rare (cf. article <strong>de</strong> Matabos et al en préparation, annexe du chapitre 5). Cette compétition<br />

sur les gammes <strong>de</strong> températures basses et le nombre limité <strong>de</strong> vestimentifères ‘survivants’<br />

lors du tarissement <strong>de</strong> la source sur les gammes <strong>de</strong> températures hautes du preferendum,<br />

<strong>pour</strong>raient ainsi induire une sensibilité forte aux changements <strong>de</strong> température et une<br />

241


Chapitre 4<br />

extinction locale supérieure <strong>pour</strong> cette espèce. Ainsi, Bengtsson (1989) a montré que la<br />

compétition interspécifique entre trois espèces <strong>de</strong> zooplancton augmentait la probabilité<br />

d’extinction locale dans un système <strong>de</strong> métapopulation. En supposant qu’un tel<br />

changement <strong>de</strong> température n’engendre pas <strong>de</strong> compétition aussi forte avec d’autres<br />

espèces <strong>pour</strong> A. pompejana et B. thermophilus, L. elevatus <strong>pour</strong>rait présenter une plus forte<br />

sensibilité que ces espèces à <strong>de</strong>s changements <strong>de</strong> température du flui<strong>de</strong> et ainsi présenter<br />

<strong>de</strong>s extinctions plus fréquentes et une structure génétique plus marquée. Cette hypothèse<br />

reste néanmoins hautement spéculative et <strong>de</strong>man<strong>de</strong> à être testée expérimentalement (e.g.<br />

exclusion compétitive sur différentes gammes <strong>de</strong> températures). De plus, d’autres facteurs,<br />

tels que la composition chimique du flui<strong>de</strong> ou la disponibilité en nourriture, <strong>pour</strong>raient<br />

également influencer la compétition interspécifique et potentiellement modifier le régime<br />

d’extinction locale (cf Matabos et al 2008a).<br />

L’hypothèse d’une plus forte sensibilité à l’extinction est également à relativiser au vu<br />

<strong>de</strong>s tailles efficaces <strong>de</strong> populations. En effet, une espèce qui présente une gran<strong>de</strong> taille<br />

efficace <strong>de</strong> populations est théoriquement moins assujettie à la dérive génétique et aurait<br />

ainsi une plus faible probabilité d’extinction qu’une espèce ayant une taille efficace plus<br />

faible (Li & Graur 1991). L’espèce L. elevatus présentent généralement <strong>de</strong>s tailles <strong>de</strong><br />

populations bien supérieures à celles <strong>de</strong> A. pompejana ou même à celle <strong>de</strong> B. thermophilus<br />

(Jollivet, communication personnelle), pouvant laisser penser que L. elevatus présente une<br />

taille efficace supérieure aux <strong>de</strong>ux autres espèces étudiées et donc, au contraire, <strong>de</strong>vrait<br />

mieux résister aux extinctions locales. Néanmoins, une gran<strong>de</strong> taille <strong>de</strong> population ne<br />

signifie pas forcément que la taille efficace l’est également. Ainsi, la taille efficace estimée<br />

sur L. fucensis (Juan <strong>de</strong> Fuca, dorsale du Pacifique Nord, Johnson et al 2006) à l’ai<strong>de</strong> d’un<br />

modèle d’Isolement avec Migration (Hey & Nielsen 2004, 2007), est du même ordre <strong>de</strong><br />

gran<strong>de</strong>ur que celle estimée par ce même modèle dans ce chapitre sur A. pompejana, mais<br />

10X plus faible que celle estimée sur B. thermophilus. Cette similitu<strong>de</strong> <strong>de</strong> taille efficace<br />

entre L. elevatus et A. pompejana <strong>pour</strong>rait s’expliquer par le mo<strong>de</strong> <strong>de</strong> reproduction <strong>de</strong> ces 2<br />

espèces. Ces 2 espèces utilisent en effet l’appariement mâle/femelle avec transmission du<br />

sperme et fécondation interne <strong>de</strong>s ovocytes <strong>pour</strong> se reproduire (Fretter 1988, Pradillon &<br />

Gaill 2003), ce qui nécessite une recherche <strong>de</strong> partenaire dans son proche entourage<br />

(reproduction entre individus proches).<br />

Une <strong>de</strong>uxième différence observée entre les 3 espèces cibles rési<strong>de</strong> dans la position<br />

géographique <strong>de</strong> la barrière la plus récente. La position géographique <strong>de</strong> la barrière ayant<br />

242


Chapitre 4<br />

engendré une vicariance concomitante entre les espèces il y a 1,3 Ma, varie en effet selon<br />

l’espèce considérée. Cette barrière est située à l’Equateur (entre 9°50’N et 7°25’S) <strong>pour</strong> A.<br />

pompejana mais elle se positionne entre 7°25’S et 14°S <strong>pour</strong> B. thermophilus (1 seul<br />

individu échantillonné à 7°25’S <strong>pour</strong> L. elevatus). De la même façon, la zone <strong>de</strong> contact<br />

entre lignées mitochondriales est plus ou moins étendue géographiquement en fonction <strong>de</strong><br />

l’espèce. Ainsi, les lignées mtCOI d’A. pompejana sont bien séparées géographiquement.<br />

Au contraire, B. thermophilus présente un mélange <strong>de</strong> ces <strong>de</strong>ux lignées divergentes au<br />

Nord <strong>de</strong> sa barrière et les espèces Nord et Sud <strong>de</strong> L. elevatus sont essentiellement en<br />

contact à 9°50’N. De plus, alors que L. elevatus et A. pompejana présente une zone<br />

d’hybridation relativement restreinte (avec <strong>de</strong>s hybri<strong>de</strong>s trouvés essentiellement à 9°50’N<br />

et 7°25’S-14°S, respectivement <strong>pour</strong> ces <strong>de</strong>ux espèces) qui s’apparente plus à la notion <strong>de</strong><br />

front d’hybridation, la zone d’hybridation <strong>de</strong> B. thermophilus s’étend à la fois au Nord et<br />

au Sud <strong>de</strong> la barrière présumée (au Nord : 2 lignées mitochondriales alors que certains<br />

gènes tels que Sulfo 1 et SAHH présentent un unique cla<strong>de</strong>, au Sud : détection d’hybri<strong>de</strong>s<br />

sur le gène SAHH <strong>de</strong> 7°25’S à 17°34’S) dans une logique plus proche <strong>de</strong> la zone<br />

d’hybridation en mosaïque décrite par Bierne et al (2003) chez le genre Mytilus. Ces<br />

différences <strong>de</strong> distribution géographique (<strong>de</strong>s lignées mitochondriales et <strong>de</strong>s hybri<strong>de</strong>s) sont<br />

en accord avec les capacités <strong>de</strong> dispersion supposées plus importantes chez B.<br />

thermophilus que chez les autres espèces et le vecteur <strong>de</strong> dispersion (courants <strong>de</strong> fond<br />

canalisés par la vallée axiale versus courants <strong>de</strong> surface). En effet, par analogie aux larves<br />

<strong>de</strong> Bathymodiolus childressi (Arellano & Young 2009), les larves planctotrophes <strong>de</strong> B.<br />

thermophilus <strong>pour</strong>raient être capables d’atteindre <strong>de</strong>s couches d’eau soumises à la lumière<br />

du Soleil, leur permettant ainsi d’accé<strong>de</strong>r à une plus gran<strong>de</strong> quantité <strong>de</strong> nourriture et passer<br />

rapi<strong>de</strong>ment d’une larve pedivéligère <strong>de</strong> 90 µm à une post-larve dont le diamètre atteint 400<br />

µm (Le Pennec 1988, Dixon et al 2007). Elles sont donc susceptibles <strong>de</strong> rester plus<br />

longtemps dans la colonne d’eau que les larves d’A. pompejana et L. elevatus, mais<br />

également <strong>de</strong> disperser en suivant <strong>de</strong>s courants <strong>de</strong> surface (potentiellement plus forts que<br />

les courants profonds et <strong>de</strong> direction généralement différente, Thomson et al 1990, Cannon<br />

& Pashinski 1997), ces <strong>de</strong>rniers étant peu ou pas du tout affectés par la mise en place <strong>de</strong><br />

barrières tectoniques telles que le décalage <strong>de</strong> la dorsale par faille transformante.<br />

243


Chapitre 4<br />

244


Chapitre 5<br />

Chapitre 5 : Conclusion, discussion générale et perspectives<br />

La distribution mais également la variabilité <strong>de</strong> l’activité <strong>de</strong>s sources<br />

hydrothermales profon<strong>de</strong>s font <strong>de</strong> cet environnement un exemple clé du concept <strong>de</strong><br />

métapopulation (Vrijenhoek 1997, Jollivet et al 1999), avec <strong>de</strong> nombreux événements<br />

d’extinctions <strong>de</strong> populations et une (re)colonisation rapi<strong>de</strong> <strong>de</strong>s sites actifs parfois distants<br />

<strong>de</strong> plusieurs centaines <strong>de</strong> kilomètres. À cela s’opposent <strong>de</strong>s conditions environnementales<br />

‘extrêmes’ (H 2 S, CO 2 , CH 4 , métaux Fe, Mn, Cu Zn, et As) relativement stables dans<br />

l’espace et perdurant <strong>de</strong>puis la création <strong>de</strong>s océans (Hannigton & Jonasson 1995), générant<br />

ainsi <strong>de</strong> fortes pressions <strong>de</strong> sélection homogénéisante. De plus, suivant le modèle <strong>de</strong><br />

dispersion adopté, les espèces <strong>pour</strong>ront présenter une homogénéisation (modèle migrant)<br />

ou <strong>de</strong> fortes différenciations génétiques entre les champs hydrothermaux<br />

géographiquement distants (modèle propagule). Outre la distance entre sites, <strong>de</strong>s barrières<br />

à la dispersion, éventuellement couplées à <strong>de</strong>s origines <strong>de</strong> colonisation différentes entre<br />

provinces géographiques (pré-compétences d’espèces phylogénétiquement indépendantes à<br />

coloniser <strong>de</strong>s environnements réduits), ont également façonné la distribution <strong>de</strong> la faune<br />

hydrothermale. Certaines barrières majeures ayant profondément impacté la composition<br />

<strong>de</strong> la faune, plusieurs provinces biogéographiques ont alors été définies (Van Dover et al<br />

2002, Tyler & Young 2003, Bachraty et al 2009). L’objectif <strong>de</strong> cette thèse visait à mieux<br />

comprendre l’histoire <strong>de</strong> la colonisation et le rôle <strong>de</strong>s failles transformantes et <strong>de</strong>s courants<br />

profonds dans la distribution actuelle <strong>de</strong>s espèces présentes le long <strong>de</strong> la dorsale du<br />

Pacifique oriental. Pour cela, <strong>de</strong>ux approches complémentaires <strong>de</strong> phylogéographie ont été<br />

réalisées : (1) une approche multi-spécifique menée sur 7 espèces en utilisant le gène<br />

mitochondrial Cytochrome Oxydase I et (2) une approche multi-locus effectuée sur 3<br />

espèces cibles A. pompejana, B. thermophilus et L. elevatus. Ces approches ont ainsi<br />

permis <strong>de</strong> dégager <strong>de</strong>s hypothèses fortes sur l’histoire évolutive <strong>de</strong> ces espèces, telles que<br />

(1) la présence d’un événement <strong>de</strong> vicariance commun entre espèces, daté <strong>de</strong> 1,3 Ma, (2)<br />

une expansion démographique récente (< 500 000 ans) <strong>de</strong> ces espèces au Sud <strong>de</strong> l’EPR et<br />

(3) la détection <strong>de</strong> zones <strong>de</strong> contact secondaire plus ou moins larges entre le Nord et le Sud<br />

<strong>de</strong> la dorsale chez les 3 espèces étudiées via l’approche multi-locus. L’approche multilocus<br />

nous a également permis <strong>de</strong> détecter <strong>de</strong>s locus soumis à divers types <strong>de</strong> sélection (i.e.<br />

balayage sélectif, sélection balancée, sélection purifiante).<br />

245


Chapitre 5<br />

1. L’EPR : une ou <strong>de</strong>ux provinces biogéographiques ?<br />

La dorsale du Pacifique oriental (EPR) fût la première dorsale considérée comme une<br />

province biogéographique, supposée séparée <strong>de</strong> la dorsale du Pacifique Nord suite à la<br />

subduction <strong>de</strong> la plaque Américaine sous la plaque Farallon (Tunnicliffe 1991), il y a<br />

environ 35 millions d’années. Cette considération en tant que province biogéographique à<br />

part entière a été étudiée sur la base <strong>de</strong> sa composition faunistique. Plusieurs étu<strong>de</strong>s ont<br />

alors défini d’autres provinces biogéographiques sur cette même base (e.g. Tunnicliffe et al<br />

1998, Van Dover et al 2002, Tyler et al 2003, Bachraty et al 2009). Si ces auteurs<br />

s’accor<strong>de</strong>nt généralement sur la définition <strong>de</strong> cinq provinces biogéographiques principales<br />

(EPR, Dorsale du Pacifique Nord, Bassins arrière-arc du Pacifique ouest, dorsale Médio<br />

Atlantique, Océan Indien), certains proposent <strong>de</strong> rassembler <strong>de</strong>ux provinces en une seule<br />

(e.g. Bassins arrière-arc du Pacifique ouest et Océan Indien : Tyler et al 2003, Bachraty et<br />

al 2009) ou <strong>de</strong> séparer une province en <strong>de</strong>ux distinctes (e.g. dorsale Médio Atlantique, Van<br />

Dover et al 2002). L’EPR a ainsi été récemment considérée comme pouvant correspondre à<br />

<strong>de</strong>ux provinces biogéographiques différentes, définies <strong>de</strong> part et d’autre <strong>de</strong> l’équateur<br />

(Bachraty et al 2009). Si en milieu terrestre les barrières à la dispersion permettent <strong>de</strong><br />

définir les provinces biogéographiques, elles correspon<strong>de</strong>nt généralement aux frontières<br />

entre continents. A l’inverse, les provinces biogéographiques marines sont plus difficiles à<br />

appréhen<strong>de</strong>r (Cox & Moore 2001), notamment du fait <strong>de</strong> l’interconnexion <strong>de</strong>s océans,<br />

l’existence d’une dimension supplémentaire : la profon<strong>de</strong>ur et <strong>de</strong> l’accessibilité moindre au<br />

milieu.<br />

Dans un travail où j’ai pris une part importante d’activité, Matabos et al (en<br />

préparation, cf. résultats présentés en annexe <strong>de</strong> ce chapitre) ont étendu le jeu <strong>de</strong> données<br />

<strong>de</strong> Bachraty et al (2009) en y ajoutant la composition <strong>de</strong>s faunes échantillonnées en 2004<br />

durant la campagne océanographique Biospeedo. L’analyse en présence/absence <strong>de</strong>s<br />

espèces le long <strong>de</strong> cette dorsale confirme ainsi l’existence d’une différence <strong>de</strong> composition<br />

<strong>de</strong> ces faunes entre l’EPR Nord et Sud et renforce l’hypothèse <strong>de</strong> <strong>de</strong>ux provinces<br />

biogéographiques <strong>de</strong> part et d’autre <strong>de</strong> l’Equateur. Ces nouvelles analyses placent la zone<br />

<strong>de</strong> séparation <strong>de</strong>s faunes en 2 points <strong>de</strong> rupture, l’équateur et la zone autour <strong>de</strong> 14°S. Ce<br />

<strong>de</strong>uxième clivage a été affiné par l’ajout <strong>de</strong>s collections récoltées pendant la campagne<br />

BioSpeedo par comparaison <strong>de</strong>s travaux <strong>de</strong> Bachraty et al (2009) qui considéraient cette<br />

246


Chapitre 5<br />

séparation à partir <strong>de</strong> 17°25’S. Les résultats obtenus en phylogéographie dans les chapitres<br />

3 et 4 montrent que cette position géographique <strong>de</strong> la séparation <strong>de</strong>s faunes correspond<br />

également au point <strong>de</strong> rupture phylogéographique le plus récent retrouvé chez la plupart<br />

<strong>de</strong>s espèces étudiées. De plus, l’analyse en ‘barco<strong>de</strong>’ du marqueur mitochondrial COI <strong>de</strong><br />

certaines espèces rares <strong>de</strong> gastéropo<strong>de</strong>s révèle l’existence <strong>de</strong> lignées mitochondriales<br />

divergentes, géographiquement structurées autour du point 14°S qui <strong>pour</strong>raient<br />

correspondre à <strong>de</strong>s espèces cryptiques, <strong>de</strong> part et d’autre <strong>de</strong> l’Equateur, renforçant<br />

également cette séparation en <strong>de</strong>ux provinces biogéographiques. Enfin, les niveaux <strong>de</strong><br />

divergence observés sur plusieurs espèces (5,5-9,5% : L. pustulosus, L. tevnianus, P. laevis<br />

et P. operculata) s’apparentent à celui trouvé entre les 2 espèces cryptiques <strong>de</strong> L. elevatus,<br />

et renforce l’hypothèse d’un premier effet vicariant plus ancien, probablement daté à 11-12<br />

millions d’années.<br />

S’il peut sembler évi<strong>de</strong>nt qu’une barrière biogéographique, engendrant <strong>de</strong>s<br />

différences <strong>de</strong> composition faunistique, puisse engendrer une forte structuration génétique<br />

<strong>de</strong>s espèces qui la traversent, une concordance parfaite entre les patrons<br />

phylogéographiques et la composition faunistique ne peut pas toujours être établie. En<br />

effet, la composition <strong>de</strong> la faune marine intertidale nord-américaine (Engle et al 1999) et la<br />

structure phylogéographique observée sur plusieurs espèces (e.g. Soltis et al 2006)<br />

semblent s’accor<strong>de</strong>r sur la côte est américaine afin <strong>de</strong> définir une barrière biogéographique<br />

entre le golfe du Mexique et la côte Atlantique <strong>de</strong> la Flori<strong>de</strong>, avec la détection d’une zone<br />

d’hybridation au niveau du Cap Canaveral <strong>pour</strong> un grand nombre d’espèces sœurs (e.g.<br />

crabes : Bert & Harrison 1988, gastéropo<strong>de</strong>s : Liu et al 1991, ou bivalves : Bert & Arnold<br />

1995). Au contraire, Burton (1998) a montré que, bien qu’engendrant un changement net<br />

<strong>de</strong>s communautés entre les provinces biogéographiques <strong>de</strong> l’Oregon et <strong>de</strong> la Californie<br />

(Doyle 1985), Point Conception ne correspondrait pas à une barrière phylogéographique<br />

intraspécifique.<br />

La concordance entre une barrière biogéographique en termes <strong>de</strong> composition<br />

faunistique et une barrière phylogéographique doit donc être considérée avec précaution.<br />

En effet, cette concordance géographique ne signifie pas forcément une concordance<br />

temporelle <strong>de</strong> l’événement <strong>de</strong> vicariance. Ainsi, Cunningham et Collins (1994) distinguent<br />

la « congruence » <strong>de</strong>s patrons <strong>de</strong> distribution, qui nécessite une concordance à la fois<br />

géographique et temporelle <strong>de</strong>s barrières, <strong>de</strong> la « pseudo-congruence » <strong>de</strong>s patrons<br />

phylogéographiques, qui correspond à une concordance géographique <strong>de</strong> ces barrières mais<br />

sur une pério<strong>de</strong> longue <strong>de</strong> colonisation du milieu, cette colonisation ayant eu lieu à <strong>de</strong>s<br />

247


Chapitre 5<br />

pério<strong>de</strong>s <strong>de</strong> temps différentes, et les barrières pouvant être issues <strong>de</strong> processus historiques<br />

différents. Donoghue & Moore (2003) proposent ainsi une classification du niveau <strong>de</strong><br />

concordance entre quatre groupes dépendant <strong>de</strong> la concordance géographique et temporelle<br />

<strong>de</strong>s événements <strong>de</strong> vicariance (Fig. 5.1). Congruence et pseudo-congruence se définissent<br />

ainsi par une correspondance géographique entre barrière et séparation <strong>de</strong>s faunes alors que<br />

congruence et incongruence correspon<strong>de</strong>nt à une concordance temporelle <strong>de</strong> l’isolement<br />

<strong>de</strong>s taxons (la pseudo-incongruence étant donc une absence <strong>de</strong> correspondance visible à la<br />

fois géographique et temporelle <strong>de</strong> l’isolement observé).<br />

Fig. 5.1 Classification <strong>de</strong> la concordance géographique et/ou<br />

temporelle entre barrières biogéographiques.<br />

Concernant l’EPR, nous avons mis en évi<strong>de</strong>nce un événement d’isolement génétique<br />

entre le Nord et le Sud <strong>de</strong> la dorsale à la fois géographiquement (si l’on admet que les<br />

différences observées entre espèces sont le reflet <strong>de</strong> capacités <strong>de</strong> dispersion différentes, <strong>de</strong><br />

réponses différentes à l’extinction locale, et/ou <strong>de</strong> barrières génétiques plus ou moins fortes<br />

entre lignées (sélection contre les hybri<strong>de</strong>s)) et temporellement concordants (1,3 Ma).<br />

Nous pouvons donc considérer que cet événement <strong>de</strong> vicariance récent a conduit à un<br />

248


Chapitre 5<br />

isolement allopatrique <strong>pour</strong> chaque espèce et reflète une barrière congruente entre ces<br />

espèces.<br />

Au contraire, l’échelle <strong>de</strong> temps ayant conduit à la différenciation <strong>de</strong>s faunes <strong>de</strong> part et<br />

d’autre <strong>de</strong> l’Equateur (concordance géographique avec la barrière intraspécifique) serait<br />

issue d’au moins un événement plus ancien <strong>de</strong> vicariance, l’estimation <strong>de</strong> la divergence<br />

entre espèces cryptiques allant <strong>de</strong> 6,5% <strong>pour</strong> L. elevatus (Matabos et al 2008b) à 12,5%<br />

<strong>pour</strong> le gastéropo<strong>de</strong> Neomphalidae Pachy<strong>de</strong>rmia laevis (Matabos et al en préparation, cf.<br />

annexe <strong>de</strong> ce chapitre). En conséquence, l’isolement intraspécifique et celui détecté sur les<br />

communautés correspon<strong>de</strong>nt, a priori, géographiquement mais pas temporellement, et<br />

peuvent ainsi être considérées comme pseudo-congruentes. Néanmoins, l’isolement<br />

intraspécifique à 1,3 Ma et celui <strong>de</strong>s communautés sont probablement issus d’évènements<br />

tectoniques distincts, le premier étant probablement lié à la présence <strong>de</strong> nombreuses failles<br />

transformantes récentes (i.e. complexe Quebrada / Discovery / Gofar, Kureth & Rea 1981,<br />

Naar & Hey 1989, Francheteau et al 1990) alors que le second <strong>pour</strong>rait être dû à la<br />

réorientation <strong>de</strong> la ri<strong>de</strong> Mathématicienne (Mammerickx & Kligtgord 1982) ou la rotation<br />

<strong>de</strong> la microplaque Bauer (Eakins & Lonsdale 2003). Cette correspondance géographique<br />

apparente <strong>pour</strong>rait ainsi n’être qu’une convergence géographique. Néanmoins, l’EPR<br />

constituerait <strong>de</strong>ux provinces biogéographiques différentes définies <strong>de</strong> part et d’autre <strong>de</strong><br />

l’Equateur.<br />

2. Connexion entre EPR et les autres provinces biogéographiques<br />

Bachraty et al (2009), en effectuant un codage présence/absence <strong>de</strong>s espèces<br />

hydrothermales échantillonnées sur 63 champs hydrothermaux (Desbruyères et al 2006),<br />

proposent que l’EPR Nord <strong>pour</strong>rait avoir constituer un rôle central dans la dispersion <strong>de</strong>s<br />

espèces hydrothermales. Ces auteurs suggèrent ainsi qu’une proportion significative <strong>de</strong> la<br />

dispersion en provenance <strong>de</strong> cette province biogéographique vers l’ensemble <strong>de</strong>s autres<br />

provinces <strong>pour</strong>rait être à l’origine <strong>de</strong> la diversité actuelle recensée sur les sources<br />

hydrothermales. Ils mettent également en évi<strong>de</strong>nce la présence d’une dispersion <strong>de</strong>s<br />

sources hydrothermales du Sud Ouest Pacifique vers le Nord Ouest Pacifique. Néanmoins,<br />

l’EPR Nord est également la province biogéographique la plus étudiée et <strong>pour</strong>rait, <strong>de</strong> ce<br />

fait, n’être pas aussi central que cela en raison du biais d’échantillonnage. Ces résultats <strong>de</strong><br />

dispersion sont donc à prendre avec précaution. Si les assemblages <strong>de</strong>s communautés <strong>de</strong><br />

l’EPR Sud sont relativement similaires à ceux <strong>de</strong> l’EPR Nord, l’observation d’espèces<br />

249


Chapitre 5<br />

typiques <strong>de</strong>s bassins arrière arc du Sud Ouest Pacifique ou <strong>de</strong> la dorsale Médio Atlantique<br />

au niveau <strong>de</strong> l’EPR Sud (et notamment 21°33S i.e. Eosipho, Shinkailepas, Chorocaris),<br />

suggérait que l’EPR Sud puisse constituer une zone <strong>de</strong> transition biogéographique entre<br />

l’EPR Nord et ces <strong>de</strong>ux autres provinces (Jollivet et al 2004). Au sein <strong>de</strong> l’EPR Sud, les<br />

champs hydrothermaux les plus au Sud (i.e. 17°25’S à 21°33’S) présentent <strong>de</strong>s<br />

compositions faunistiques différentes <strong>de</strong>s sites situés plus au Nord (i.e. 7°25’S et 14°S) et<br />

une diversité taxonomique supérieure (Matabos et al en préparation, annexe <strong>de</strong> ce<br />

chapitre), suggérant que les champs 17°25’S à 21°33’S <strong>pour</strong>raient effectivement constituer<br />

une zone <strong>de</strong> transition <strong>de</strong>s faunes entre l’EPR Nord et d’autres provinces biogéographiques<br />

telles que le Sud Ouest Pacifique. Il semblerait donc que la dispersion <strong>de</strong>s faunes <strong>de</strong> l’EPR<br />

Nord (si on accepte l’hypothèse d’un rôle central <strong>de</strong> l’EPR Nord dans la colonisation <strong>de</strong>s<br />

sources par la faune mo<strong>de</strong>rne actuelle) <strong>pour</strong>rait également avoir transité via l’EPR Sud<br />

<strong>pour</strong> atteindre les autres provinces biogéographiques.<br />

3. Barrières semi-perméables : rôle <strong>de</strong>s failles transformantes et <strong>de</strong>s<br />

courants divergents ?<br />

Si ce travail <strong>de</strong> thèse nous a permis <strong>de</strong> mettre en évi<strong>de</strong>nce l’importance <strong>de</strong>s failles<br />

transfomantes et <strong>de</strong>s courants profonds en tant que barrière aux flux <strong>de</strong> gènes, l’approche<br />

multi-locus sur trois espèces cibles a montré le caractère semi-perméable <strong>de</strong> ces barrières,<br />

avec la détection d’individus hybri<strong>de</strong>s ou introgressés. Cependant, il est difficile <strong>de</strong> savoir<br />

<strong>de</strong> quand date cette semi-perméabilité et donc <strong>de</strong> conclure quant à un isolement<br />

allopatrique en cours avec migration (modèle d’Isolement avec Migration, Nielsen &<br />

Wakeley 2001) ou une remise en contact secondaire <strong>de</strong> lignées divergentes (Nosil et al<br />

2009). En effet, le modèle d’Isolement avec Migration suggère une migration entre lignées<br />

relativement anciennes en situation parapatrique. Dans notre cas, à la vue du mécanisme<br />

très progressif <strong>de</strong> séparation <strong>de</strong>s segments d’une dorsale (formation d’un décalage lent et<br />

continue par un effet <strong>de</strong> coulissage donnant lieu à la faille transformante), il serait sans<br />

doute intéressant <strong>de</strong> considérer que le mo<strong>de</strong> <strong>de</strong> spéciation prévalent correspond à une<br />

situation d’allopatrie admettant une cessation très progressive <strong>de</strong>s flux <strong>de</strong> gènes semblable<br />

au modèle décrit par Hey et Wakeley au moins pendant les 2 à 3 premiers millions<br />

d’années <strong>de</strong> l’isolement. Chez A. pompejana et B. thermophilus, tout concor<strong>de</strong> ainsi à<br />

montrer que l’isolement est actuellement en cours. Ce scénario est également soutenu par<br />

250


Chapitre 5<br />

le non-échantillonnage <strong>de</strong> l’in<strong>de</strong>l <strong>de</strong> 1 pb au Nord <strong>de</strong> l’EPR sur le gène Lysozyme pouvant<br />

signifier que cet in<strong>de</strong>l récemment apparu au Sud est désormais bloqué par la barrière aux<br />

flux <strong>de</strong> gènes (mais l’absence <strong>de</strong> cet in<strong>de</strong>l au Nord serait à vérifier par génotypage). Au<br />

contraire, la détection d’individus introgressés (entre lignées divergentes) détectés en<br />

comparant directement la lignée mitochondriale et le ‘background’ nucléaires <strong>de</strong>s lignées<br />

ou via <strong>de</strong>s génotypages RFLP sur les gènes PGM (A. pompejana) et SAHH (B.<br />

thermophilus), ainsi que la présence d’individus introgressés à double mitochondries chez<br />

L. elevatus suggèrent plutôt une remise en contact secondaire <strong>de</strong> ces lignées divergentes :<br />

une situation relativement logique chez L. elevatus ou le processus <strong>de</strong> spéciation est engagé<br />

<strong>de</strong>puis beaucoup plus longtemps et <strong>pour</strong> lequel une zone <strong>de</strong> tension est manifeste à 9°50N.<br />

4. Perspectives <strong>de</strong> la thèse<br />

Ce travail <strong>de</strong> thèse nous a permis <strong>de</strong> dégager différentes perspectives, tant d’un point<br />

<strong>de</strong> vu <strong>de</strong> l’évolution <strong>de</strong>s espèces hydrothermales et <strong>de</strong> la compréhension du milieu<br />

hydrothermal, que <strong>de</strong> l’évolution <strong>de</strong>s gènes étudiés.<br />

4.1. Évolution <strong>de</strong>s espèces hydrothermales<br />

La plupart <strong>de</strong>s logiciels utilisant les modèles <strong>de</strong> coalescence ne prenant pas en compte<br />

la recombinaison comme force <strong>de</strong> diversité allélique, les analyses effectuées durant cette<br />

thèse ont, <strong>pour</strong> la plupart, été réalisées en éliminant les séquences recombinantes <strong>de</strong>s jeux<br />

<strong>de</strong> données. Or, certains logiciels récents basés sur <strong>de</strong>s métho<strong>de</strong>s Bayésiennes permettent<br />

<strong>de</strong> prendre en compte la recombinaison si celle-ci n’est pas trop élevée (e.g. PopABC,<br />

MIMAR). Ils supposent néanmoins <strong>de</strong> fournir un certain nombre <strong>de</strong> paramètres initiaux<br />

(prior) pas toujours aisés à définir lorsque les caractéristiques biologiques <strong>de</strong>s espèces sont<br />

peu connues (comme c’est le cas ici, e.g. estimation a priori <strong>de</strong> la taille efficace <strong>de</strong>s<br />

populations). Néanmoins, ces logiciels n’en <strong>de</strong>meurent pas moins intéressants et peuvent<br />

permettre <strong>de</strong> complémenter nos analyses, notamment <strong>pour</strong> les espèces cryptiques <strong>de</strong> L.<br />

elevatus <strong>pour</strong> lesquelles la recombinaison entre allèles était trop importante <strong>pour</strong> être<br />

totalement supprimée <strong>de</strong>s jeux <strong>de</strong> données. Pour ces espèces, certaines manipulations<br />

complémentaires basées sur le génotypage par RFLP <strong>pour</strong>raient également apporter <strong>de</strong>s<br />

251


Chapitre 5<br />

informations précieuses en permettant <strong>de</strong> quantifier plus finement l’hybridation le long <strong>de</strong><br />

l’EPR, à l’instar <strong>de</strong>s gènes PGM d’A. pompejana et SAHH <strong>de</strong> B. thermophilus. Ainsi, un<br />

site <strong>de</strong> restriction <strong>pour</strong> l’enzyme Mse1 a été détecté sur le gène Lysozyme chez L. elevatus<br />

en liaison avec une mutation fixée discriminant les <strong>de</strong>ux cla<strong>de</strong>s et <strong>pour</strong>rait ainsi permettre<br />

<strong>de</strong> mieux évaluer l’étendue <strong>de</strong> la zone d’hybridation entre ces espèces.<br />

La mise en évi<strong>de</strong>nce d’une différenciation génétique entre populations variant selon la<br />

latitu<strong>de</strong> et les espèces étudiées pose la question <strong>de</strong> la raison <strong>de</strong> ces disparités. Nous avons<br />

proposé dans les chapitres 3 et 4, <strong>de</strong>s hypothèses qui <strong>pour</strong>raient expliquer ces différences.<br />

Néanmoins, les connaissances biologiques (e.g. cycle <strong>de</strong> vie, capacités <strong>de</strong> dispersion,…)<br />

mais également écologiques (e.g. compétition intra et interspécifique, niveau <strong>de</strong><br />

fragmentation <strong>de</strong> l’habitat associé à l’espèce, adaptation à l’environnement local,…) <strong>de</strong> ces<br />

espèces restent toutefois limitées. Afin <strong>de</strong> mieux comprendre l’évolution <strong>de</strong>s espèces, il<br />

semble nécessaire <strong>de</strong> mieux appréhen<strong>de</strong>r leur capacité actuelle à coloniser <strong>de</strong> nouveaux<br />

sites et leur sensibilité vis-à-vis <strong>de</strong>s modifications biotiques (relations entre espèces<br />

sympatriques) et abiotiques (e.g. température, composition du flui<strong>de</strong> hydrothermal).<br />

Certaines espèces n’ont été que très peu étudiées et leur cycle <strong>de</strong> vie reste extrêmement peu<br />

connu, c’est le cas notamment du polychaete Hesiolyra bergi ou du gastéropo<strong>de</strong><br />

Eulepetopsis vitrae. Des observations biologiques directes <strong>pour</strong>raient ainsi permettre<br />

d’apporter <strong>de</strong>s informations quant à son mo<strong>de</strong> <strong>de</strong> reproduction, par exemple. Outre les<br />

observations directes, <strong>de</strong>s approches indirectes peuvent permettre d’évaluer les capacités<br />

<strong>de</strong> dispersion <strong>de</strong> ces espèces, via l’étu<strong>de</strong> <strong>de</strong> la connectivité entre populations. De telles<br />

étu<strong>de</strong>s ont déjà été engagées au niveau <strong>de</strong> l’EPR nord en modélisant la circulation <strong>de</strong>s<br />

masses d’eau au niveau <strong>de</strong> la dorsale à l’ai<strong>de</strong> <strong>de</strong> mouillage <strong>de</strong> courantomètres, <strong>de</strong> piégeage<br />

<strong>de</strong> larves dans la colonne d’eau et la plume hydrothermale ou la mise en place <strong>de</strong> modules<br />

<strong>de</strong> colonisation sur <strong>de</strong>s séries à long terme (cf. Mullineaux et al 1998). L’utilisation <strong>de</strong><br />

marqueurs génétiques suffisamment polymorphes (e.g. microsatellites, AFLP) <strong>pour</strong>rait<br />

également être extrêmement informatif et permettre, par exemple, d’assigner <strong>de</strong>s individus<br />

migrants à leur population d’origine, si une différenciation suffisamment forte entre<br />

populations est révélée. Une autre manière d’observer <strong>de</strong> manière indirecte la connectivité<br />

entre populations serait une étu<strong>de</strong> <strong>de</strong> la géochimie <strong>de</strong>s coquilles chez les bivalves ou les<br />

gastéropo<strong>de</strong>s. En effet, le flui<strong>de</strong> hydrothermal peut présenter <strong>de</strong>s compositions chimiques<br />

légèrement différentes entre champs hydrothermaux. Les coquilles <strong>de</strong>s mollusques étant<br />

élaborées en utilisant les composés chimiques présents dans le milieu lors <strong>de</strong> leur<br />

252


Chapitre 5<br />

formation, la composition chimique <strong>de</strong>s coquilles larvaires (encore présentes chez l’adulte)<br />

peut ainsi renseigner sur la provenance <strong>de</strong> chaque individu, si tant est que cette coquille<br />

larvaire se soit formée au contact du flui<strong>de</strong> hydrothermal et pas dans la colonne d’eau entre<br />

ces sites (un cas envisageable chez les gastéropo<strong>de</strong>s Lepetodrilidae mais peu vraisemblable<br />

chez les Bathymodiolinae ou les larves sont directement émises dans la colonne d’eau en<br />

périphérie <strong>de</strong>s points d’émission). Ce genre d’approche géochimique a déjà été réalisée<br />

avec succès sur <strong>de</strong>s coquilles <strong>de</strong> bivalves côtiers (Mytilus californianus, M. edulis, Becker<br />

et al 2005, 2007), <strong>de</strong> gastéropo<strong>de</strong>s intertidaux (Concholepas concholepas : Zacherl et al<br />

2003, Kelletia kelletii : Zacherl et al 2005) ou sur <strong>de</strong>s otholithes <strong>de</strong> poissons (Thorrold et al<br />

2007). Les gastéropo<strong>de</strong>s qui incubent leurs larves dans leur cavité palléale (larves formées<br />

sur le champ hydrothermal d’origine), tels que les Lepetodrilus, <strong>pour</strong>raient ainsi constituer<br />

<strong>de</strong> bons candidats à une approche géochimique <strong>de</strong> connectivité <strong>de</strong>s populations.<br />

Une meilleure compréhension du milieu profond lui-même, par une investigation <strong>de</strong>s<br />

courants profonds par exemple, <strong>pour</strong>rait également permettre <strong>de</strong> mieux expliquer les<br />

patrons <strong>de</strong> distribution <strong>de</strong>s espèces. De plus, la découverte récente d’un champs<br />

hydrothermal actif à 4°S sur la dorsale EPR avec une faune hydrothermale associée,<br />

<strong>pour</strong>rait permettre <strong>de</strong> préciser la position <strong>de</strong> la barrière géographique entre l’EPR Nord et<br />

l’EPR Sud.<br />

4.2. Évolution <strong>de</strong>s gènes<br />

4.2.1. Le gène mtCOI<br />

L’étu<strong>de</strong> du gène mitochondrial <strong>de</strong> la Cytochrome Oxydase I a permis <strong>de</strong> mettre en<br />

évi<strong>de</strong>nce un co<strong>de</strong> génétique particulier chez le polychète A. pompejana avec la méthionine<br />

qui serait codée par <strong>de</strong>ux codons différents ATG et ATA (ce <strong>de</strong>rnier codant habituellement<br />

<strong>pour</strong> l’isoleucine). Il serait extrêmement intéressant <strong>de</strong> <strong>pour</strong>suivre cette étu<strong>de</strong> du co<strong>de</strong><br />

génétique mitochondrial <strong>de</strong>s Alvinellidae en séquençant d’autres gènes mitochondriaux<br />

afin <strong>de</strong> vérifier si ce co<strong>de</strong> génétique est partagée par d’autres espèces d’Alvinellidae, voire<br />

d’autres annéli<strong>de</strong>s. D’ores et déjà, une étu<strong>de</strong> <strong>de</strong>s génomes mitochondriaux <strong>de</strong> plusieurs<br />

annéli<strong>de</strong>s dont Riftia pachyptila, Alvinella pompejana et Paralvinella grasslei est possible.<br />

Si <strong>de</strong>s séquences <strong>de</strong> ce gène ont été obtenues chez A. caudata, le nombre <strong>de</strong> séquences<br />

253


Chapitre 5<br />

effectuées restaient cependant trop faibles <strong>pour</strong> pouvoir avoir un accès réel au<br />

polymorphisme. La raison <strong>de</strong> ce faible nombre <strong>de</strong> séquences obtenues sur A. caudata est<br />

également très intéressante. En effet, sur 48 individus séquencés en direct, seule une<br />

dizaine <strong>de</strong> séquences présentaient <strong>de</strong>s chromatogrammes interprétables. Or, une fois ces<br />

séquences alignées, nous avons détecté trois types <strong>de</strong> séquences différentes. Les <strong>de</strong>ux<br />

premiers types étaient différents par <strong>de</strong>s substitutions synonymes, suggérant la présence <strong>de</strong><br />

<strong>de</strong>ux lignées mitochondriales (voire espèces cryptiques). Le troisième type <strong>de</strong> séquences<br />

était beaucoup plus intrigant puisqu’il différait <strong>de</strong>s <strong>de</strong>ux autres par la présence d’un in<strong>de</strong>l<br />

affectant le cadre <strong>de</strong> lecture <strong>de</strong> ce gène. Ainsi, la présence <strong>de</strong> cet in<strong>de</strong>l sur le gène mtCOI<br />

laisse penser à une possible duplication du gène qui <strong>pour</strong>rait ou non être passé dans le<br />

génome nucléaire. Ce genre <strong>de</strong> transfert à déjà été décrit dans la littérature sous le nom <strong>de</strong><br />

numts en analysant les génomes <strong>de</strong> certaines espèces modèles (e.g. coléoptère Tribolium<br />

ou l’abeille Apis, Pamilo et al 2007). Le nombre important <strong>de</strong> chromatogrammes illisibles<br />

sur le mtCOI <strong>de</strong> cette espèce <strong>pour</strong>rait être dû à la co-amplification <strong>de</strong> ces gènes dupliqués<br />

ou d’individus présentant une hétéroplasmie mitochondriale (cf. hybri<strong>de</strong>s <strong>de</strong> L. elevatus).<br />

De plus amples étu<strong>de</strong>s <strong>pour</strong>raient ainsi permettre <strong>de</strong> mieux comprendre l’évolution <strong>de</strong> ce<br />

gène chez A. caudata. Tout d’abord, le clonage <strong>de</strong>s produits <strong>de</strong> PCR <strong>de</strong>s individus<br />

présentant <strong>de</strong>s chromatogrammes illisibles <strong>pour</strong>raient permettre <strong>de</strong> savoir si la coamplification<br />

<strong>de</strong> plusieurs lignées paralogues est bien à l’origine <strong>de</strong> ces chromatogrammes<br />

illisibles. Ensuite, une extraction séparée <strong>de</strong> l’ADN mitochondrial et l’ADN génomique<br />

suivie d’une amplification par PCR sur ces <strong>de</strong>ux types d’ADN <strong>pour</strong>rait permettre <strong>de</strong> savoir<br />

si une duplication <strong>de</strong> ce gène est passée dans le génome nucléaire ou si cette duplication<br />

potentielle s’est effectuée dans le génome mitochondrial.<br />

La détection d’individus présentant une double hérédité parentale et étant hybri<strong>de</strong>s<br />

entre les <strong>de</strong>ux espèces cryptiques <strong>de</strong> L. elevatus, est également un point intéressant à<br />

creuser. Cependant, elle nécessite d’effectuer un échantillonnage adapté <strong>de</strong> L. elevatus,<br />

notamment à 9°50’N, d’en déterminer le sexe puis d’examiner la double hérédité en<br />

fonction du sexe <strong>de</strong>s individus. Si un nombre suffisant d’individus à double hérédité<br />

parentale est détecté, il serait alors possible <strong>de</strong> regar<strong>de</strong>r si ces individus présentent un sexe<br />

particulier (e.g. si seuls les mâles ont <strong>de</strong>ux mitochondries comme chez les Mytilus : Zouros<br />

et al 1994). Examiner le caractère <strong>de</strong> double hérédité <strong>de</strong>s mitochondries chez les larves ou<br />

même les œufs fécondés, <strong>pour</strong>rait également apporter une information importante, à<br />

savoir : si la mitochondrie du père (transmise par le sperme) est transmise à la <strong>de</strong>scendance<br />

254


Chapitre 5<br />

puis perdue au cours du développement ou <strong>de</strong> la ségrégation méiotique. L’existence<br />

d’individus probablement F1 à double mitochondries laisse supposer en effet que les<br />

mécanismes pré-zygotiques n’empêchent pas la double hérédité parentale. Néanmoins, il<br />

va sans dire que ces questions dépen<strong>de</strong>nt <strong>de</strong> l’échantillonnage et présentent actuellement<br />

un véritable défi.<br />

4.2.2. Phosphoglucomutase d’A. pompejana<br />

La comparaison allozymes/RFLP sur le gène PGM d’A. pompejana a permis <strong>de</strong><br />

détecter la mutation non synonyme probablement responsable <strong>de</strong> la différence <strong>de</strong> charge<br />

entre allèles 90 et 100 sur les individus du Sud <strong>de</strong> l’EPR. Néanmoins, cette mutation<br />

n’explique pas cette différence sur les individus du Nord <strong>de</strong> l’EPR, indiquant la présence<br />

d’une homoplasie <strong>de</strong> charge. Le séquençage <strong>de</strong> la totalité du gène <strong>de</strong> la PGM montre qu’un<br />

seul remplacement non-synonyme Q/E en fréquence intermédiaire explique le<br />

polymorphisme <strong>de</strong> la lignée nord, le site étant différent <strong>de</strong> celui trouvé dans la lignée sud<br />

(D. Jollivet, données non publiées). Il serait donc extrêmement intéressant <strong>de</strong> cibler une<br />

étu<strong>de</strong> sur ce gène au voisinage <strong>de</strong> la (ou <strong>de</strong>s) mutation(s) responsable(s) du changement <strong>de</strong><br />

charge et dans les parties exoniques les plus éloignées <strong>de</strong> cette mutation, dans les 2 lignées<br />

d’Alvinella pompejana. Ce gène étant supposé sous sélection balancée, tout du moins au<br />

Nord <strong>de</strong> la dorsale, cette mutation <strong>de</strong>vrait être accompagnée d’un excès <strong>de</strong> diversité<br />

nucléotidique à son voisinage dans l’hypothèse d’une sélection balancée (cf. Filatov &<br />

Charlesworth 1999). Une telle étu<strong>de</strong> par fenêtre coulissante sur la diversité du gène n’a<br />

malheureusement pas été menée sur la lignée sud d’Alvinella pompejana car la mutation<br />

Q/E incriminée se trouvait à l’une <strong>de</strong>s extrémités du fragment d’ADN amplifié.<br />

255


Chapitre 5<br />

5. Annexe du chapitre 5<br />

Faunal changes and geographic crypticism indicate the occurrence of a biogeographic<br />

transition zone along the Southern East-Pacific Rise (SEPR)<br />

Matabos M. 1,2* , Plouviez S. 2,3 , Hour<strong>de</strong>z S. 2,3 , Desbruyères D. 4 , Legendre P. 5 , Warén A. 6 ,<br />

Jollivet D. 2,3 and Thiébaut E. 2,3 ,<br />

1 Muséum National d’Histoire Naturelle, Département Milieux et Peuplements<br />

Aquatiques, UMR BOREA 7208 (MNHN, UPMC, CNRS), CP 53, 61 rue Buffon, F-<br />

75231 Paris ce<strong>de</strong>x 05, France<br />

2<br />

CNRS, UMR 7144, BP 74, F-29682 <strong>Roscoff</strong> ce<strong>de</strong>x, France<br />

3 Université Pierre et Marie Curie-Paris 6, UMR 7144, <strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>,<br />

Laboratoire Adaptation et Diversité en Milieu Marin, BP74, F-29682 <strong>Roscoff</strong> ce<strong>de</strong>x,<br />

France<br />

4 Ifremer, Département Etu<strong>de</strong> <strong>de</strong>s Ecosystèmes Profonds, BP70, F-29280 Plouzané,<br />

France<br />

5<br />

Département <strong>de</strong> sciences biologiques, Université <strong>de</strong> Montréal, CP 6128, Succursale<br />

Centre-ville, Montréal, Québec, Canada H3C 3J7<br />

6 Swedish Museum of Natural History, Box 5007, SE-10405, Stockholm, Swe<strong>de</strong>n<br />

*corresponding author: mmatabos@uvic.ca<br />

Present address: School of Earth and Ocean Sciences, University of Victoria, PO Box<br />

3065, STN CSC, Victoria, BC V8W 3V6, Canada<br />

256


Chapitre 5<br />

ABSTRACT.<br />

Aim. Following their discovery in 1977 during the exploration of the Galapagos Spreading<br />

Center, <strong>de</strong>ep-sea hyrothermal vents have been now reported along all active mid-ocean<br />

ridges and back-arc basins spreading centers. Although 5 to 7 distinct biogeographic<br />

provinces have been then <strong>de</strong>scribed, <strong>de</strong>limitations between biogeographic entities remain<br />

questionable due to methodological issues. Here we examine biogeographic patterns of the<br />

hydrothermal vent fauna along the East Pacific Rise (EPR) and <strong>de</strong>termine the relative role<br />

of regional and local factors on the distribution of biodiversity associated with mussel beds<br />

along a poorly explored zone, the southern EPR.<br />

Location. East Pacific Rise<br />

Methods. Species list along the EPR was compiled from the literature and completed with<br />

data recovered during the last French cruise BIOSPEEDO, held in 2004 along the southern<br />

Eastern Pacific. Biogeographic patterns were assessed by combining i<strong>de</strong>ntification of<br />

morphological species and a molecular barcoding approach. The i<strong>de</strong>ntification of a<br />

geographic break was performed using a multivariate regression tree. The spatial variation<br />

of the community structure in the mussel beds of SEPR in relation to environmental factors<br />

was analysed through the measure of different diversity indices, a cluster analysis and a<br />

redundancy canonical analysis.<br />

Results. Sequencing of the Cytochrome Oxidase I gene revealed the occurrence of several<br />

cryptic species complexes along the EPR separating southern and northern cla<strong>de</strong>s on both<br />

part of the Equator. Furthermore, the Biospeedo cruise led to the discovery of at least 10<br />

new species between 7°S and 21°S. The resulting shift in community structure highlighted<br />

by the MRT analysis was located at 14°S and 17°25S suggesting that the southern part of<br />

the SEPR (17°-21°S) constitutes a biogeographic transition zone in the vent fauna along<br />

the EPR. At regional scale, the latitu<strong>de</strong> combined with the type of venting sampled had a<br />

significant impact on community structure associated with mussel beds along the SEPR.<br />

However those factors were not sufficient to explain variation in species composition<br />

suggesting a strong influence of other local factors such as small-scale habitat<br />

heterogeneity and/or biotic interactions.<br />

257


Chapitre 5<br />

Keywords East Pacific Rise, mussel beds, hydrothermal vents, species richness,<br />

biodiversity, biogeographic patterns, community structure, barcoding.<br />

Running title. Biogeographic transition zone along the south EPR<br />

258


Chapitre 5<br />

INTRODUCTION<br />

The discovery of <strong>de</strong>nse communities of en<strong>de</strong>mic species associated with hydrothermal<br />

vents on the Galapagos Rift in 1977 offered a new perception of <strong>de</strong>ep-sea benthic fauna,<br />

far from what was commonly accepted in the vast poorly-productive abyssal plain<br />

(Lonsdale, 1977). Hydrothermal vents are discontinuously distributed in oases along<br />

oceanic ridges and characterized by an ephemeral lifespan vent fauna related to tectonic<br />

events and volcanic activities occurring along the ridge crest. The associated marine<br />

communities harbour high <strong>de</strong>nsities of macrobenthic fauna along with low species richness<br />

and low diversity resulting from the overwhelming numerical dominance of a few<br />

symbiotic species. They are clustered around vents where they thrive on microbial<br />

chemoautotrophic primary production (Van Dover, 2000). These oases are characterized<br />

by a low number of rare species which are heterogeneously distributed in time and space<br />

<strong>de</strong>pending on the vent conditions and faunal succession occurring during the early<br />

colonisation of new vent emissions (Jollivet, 1996; Shank et al. 1998). The vent fauna has<br />

adapted to the great instability of the habitat where strong variations of the geological<br />

setting (in relation with the magmatic and/or tectonic activity) and the physical-chemical<br />

conditions (related to the temporal dynamic of the fluid discharge) naturally occur on a<br />

broad range of spatial (centimetres to thousands kilometres) and temporal (seconds to<br />

years) scales (Juniper & Tunnicliffe, 1997; Sarrazin et al., 1997).<br />

The en<strong>de</strong>mism at the specific and generic levels has raised questions about vent<br />

species origin and evolution: it was suggested that they originated and then in<strong>de</strong>pen<strong>de</strong>ntly<br />

evolved from late Cretaceous and early Tertiary, around 65 Ma (Little & Vrijenhoek,<br />

2003). But the colonization pathways and the links between the hydrothermal systems in<br />

the global ocean are still poorly un<strong>de</strong>rstood <strong>de</strong>spite the growing literature addressing<br />

biogeographic issues during the last <strong>de</strong>ca<strong>de</strong> (e.g. Tunnicliffe & Fowler, 1996; Tunnicliffe<br />

et al., 1998; Tyler et al., 2003; Desbruyères et al. 2006a; Bachraty et al., 2009). The<br />

present distribution of vent fauna mainly reflects historical events (i.e. plate tectonics) and<br />

species abilities to colonize distant territories (i.e. dispersion and migration) (review in<br />

Tunnicliffe et al., 1998). While dispersal and migration are major processes linking<br />

neighbouring vent faunas (Tunnicliffe & Fowler, 1996), the interval between vents<br />

<strong>de</strong>pends on the geodynamic nature of the ridge (i.e. volcanism or tectonism).<br />

Based on faunal composition, the global ocean ridge system consists in a nearly<br />

continuous ca. 60000 km volcanic chain representing five main biogeographic provinces:<br />

the Northeast Pacific, the East Pacific Rise (EPR), the Northwestern Pacific back arc<br />

259


Chapitre 5<br />

basins, the Mid-Atlantic ridge (MAR), and the Southwest Pacific and Indian Ocean.<br />

However, <strong>de</strong>pending on the authors, five to seven biogeographic entities can be erected<br />

from the literature often differing in <strong>de</strong>limitations within an ocean (Tunnicliffe, 1991;<br />

Tunnicliffe et al., 1998; Van Dover et al., 2002; Tyler et al., 2003; Bachraty et al., 2009).<br />

As an example, while the Mid-Atlantic ridge is consi<strong>de</strong>red as a single biogeographic<br />

province by most authors (e.g. Tunnicliffe et al., 1998; Desbruyères et al., 2006a; Tyler et<br />

al., 2003; Bachraty et al., 2009), Van Dover et al., (2002) <strong>de</strong>fined the Azores plateau and<br />

the MAR as two distinct provinces. Similarly, <strong>de</strong>pending on the authors, the Western<br />

Pacific back-arc basins (BAB) comprise one or two (north vs south) biogeographic<br />

provinces (e.g. Tyler et al., 2003; Bachraty et al., 2009). The differences among authors<br />

<strong>de</strong>pend on the methodology and criteria used to <strong>de</strong>fine the biogeographic units, as no clear<br />

comprehensive classification exists for the <strong>de</strong>ep-sea (Spalding et al., 2007), and the global<br />

database of species inventory is far from being complete and still growing with years.<br />

The specialized nature of vents and their geographical isolation on active spreading<br />

centres are interesting features to study vicariance, species dispersal and colonisation<br />

pathways (Tunnicliffe, 1988). Two complementary approaches are commonly used in<br />

biogeography: historical and ecological biogeography. While ecological biogeography<br />

investigates the current mechanisms of communication and dispersion at the regional scale,<br />

historical biogeography studies tectonic/climatic events that have lead to the creation or<br />

dislocation of geographic barriers. When consi<strong>de</strong>ring oceanic ridges at a global scale,<br />

tectonic events, creating physical barriers such as transform faults or microplates, can lead<br />

to the isolation of species in allopatry. Geographic isolation can yield pairs of vicariant<br />

species evolving in<strong>de</strong>pen<strong>de</strong>ntly, but filling the same ecological niche. This process has<br />

been suggested to be an important driving force of vent fauna evolution (Tunnicliffe &<br />

Fowler, 1996; Tunnicliffe et al., 1998) and appears to be more and more accepted for<br />

species in most marine taxa (Heads, 2005). Ecological biogeography studies the factors<br />

that <strong>de</strong>fine the spatial distributions of species in the present time (Monge-Najera, 2008),<br />

thus processes acting at regional to local scales. At a regional oceanic scale, topographic<br />

features, geodynamic environment and hydrodynamism highly influence the<br />

extinction/recolonisation rate of the fauna, which, in turn, govern metapopulation<br />

processes by creating gaps in species distributions (e.g. Jollivet et al. 1999, Won et al.<br />

2003). Those processes are likely to affect the regional pool of species. At a local scale (i.e.<br />

vent fields), community attributes can be highly variable within a site, along the fluid<br />

mixing gradient (Sarrazin et al., 1997; Sarrazin & Juniper, 1999; Sarrazin et al., 1999), and<br />

260


Chapitre 5<br />

between sites according to the successional stage or age of the site, which is related to the<br />

temporal dynamics of vents resulting from the waning and waxing of hydrothermal fluid<br />

discharge (Shank et al., 1998; Sarrazin et al., 1999; Mullineaux et al., 2000). Biotic<br />

factors, such as facilitation, predation, competition for food or space, by controlling<br />

recruitment and abundances, may play a significant role in structuring local benthic<br />

communities (Mullineaux et al., 1998; Micheli et al., 2002; Mullineaux et al., 2003;<br />

Sancho et al., 2005). Finally, spatial habitat heterogeneity, related to spatial variations in<br />

fluid fluxes, results in the occurrence of multiple microhabitats at small scale (Sarrazin et<br />

al., 1997; Mills et al. 2007) and could promote sympatric speciation. This process, also<br />

called ecological speciation, has been suggested to fix divergence c.a. 20 times faster than<br />

species formation in allopatry but is rarely observed in natura (Via, 2001; Briggs, 2006).<br />

Thus, local factors do not only act at the community level but can also have broa<strong>de</strong>r<br />

significance. Ecology and biogeography are thus two complementary approaches in or<strong>de</strong>r<br />

to un<strong>de</strong>rstand biodiversity patterns (Briggs, 2007), ecological processes acting at different<br />

spatial and temporal scales being often imbricate (Levin et al., 2001). On the other hand,<br />

local processes influencing communities structure and biodiversity patterns can blur<br />

broa<strong>de</strong>r biogeographical patterns if a non-adapted sampling strategy is un<strong>de</strong>rtaken (e.g.<br />

lack of samples and/or replicates compared to the scope of the study).<br />

Since 1977, taxonomists have <strong>de</strong>scribed 533 morphological species comprised in<br />

320 genera from vents (Desbruyères et al., 2006b). However, all known vent sites did not<br />

un<strong>de</strong>rgo the same sampling effort, and biodiversity may be un<strong>de</strong>restimated for poorly<br />

explored zones. Moreover, previous biogeographic analyses did not take advantage of the<br />

occurrence of sibling species, which is wi<strong>de</strong>spread in marine environment (Knowlton,<br />

1993). Ad<strong>de</strong>d to the i<strong>de</strong>ntification of morphospecies, molecular systematics is a powerful<br />

tool to address biogeographic issues as it tells how species separate or expand in time and<br />

space. Comparative phylogeography (i.e. comparison of phylogeographic patterns among<br />

multiple co-distributed taxa) has strong parallel with historical biogeography and can help<br />

in i<strong>de</strong>ntifying vicariant events unveiled by the taxonomy (Arbogast & Kenagy, 2001). To<br />

date, cryptic species have been reported from hydrothermal vents for many taxa separated<br />

by physical geological discontinuities that offset the ridge system (e.g. Won et al., 2003;<br />

Hurtado et al., 2004; Young et al. 2008). Along the EPR, the presence of cryptic species<br />

complexes was highlighted for lepetodrilid species (Johnson et al., 2008; Matabos et al.,<br />

2008b). A strong genetic divergence between a northern and a southern lineages was also<br />

observed from a set of taxonomic groups suggesting the occurrence of permeable genetic<br />

261


Chapitre 5<br />

boundaries around the Equator (e.g. Hurtado et al., 2002; Won et al. 2003, Hurtado et al.,<br />

2004; Matabos et al., 2008b). This hypothesis was recently validated by Plouviez et al.<br />

(2009) who <strong>de</strong>monstrated that the recent ridge history of the East Pacific Rise led to a<br />

vicariant event that induced genetic divergence across the Equator about 1.3 millions years<br />

ago for nearly all the vent taxa.<br />

The Southern East-Pacific Rise (SEPR) is among the fastest spreading ridge centres<br />

in the world (150 mm.an -1 ); it is characterized by a high instability in hydrothermal venting<br />

often associated with recent effusive volcanism (Auzen<strong>de</strong> et al., 1994; Fouquet et al.,<br />

1994; Embley et al., 1998). While this portion of the EPR has been extensively studied<br />

from the geological, tectonic and geochemical points of view (Auzen<strong>de</strong> et al., 1994;<br />

Fouquet et al., 1994; Fornari & Embley, 1995; Fujioka et al., 1995; Embley et al., 1998),<br />

few papers are <strong>de</strong>aling with faunal composition across vent fields (Van Dover, 2002) and<br />

the only studies found in the literature were based on few samples and study sites. Thus<br />

first analyses of faunal assemblages along the SEPR <strong>de</strong>scribed fauna similar to the<br />

northern part of the EPR (Geistdoerfer et al., 1995; Halanych et al., 1999). From<br />

quantitative samples, (Van Dover, 2002, 2003), proposed that the East Pacific Rise<br />

represents a single hydrothermal biogeographic province over at least 31 <strong>de</strong>grees of<br />

latitu<strong>de</strong>, differences relying solely on the relative abundances of the dominant species and<br />

the distribution of rare species. This latter author suggested that fields are not far enough to<br />

prevent larval and/or adult dispersal in a stepping-stone manner (Van Dover, 2002).<br />

However, the author came to these conclusions on the analysis of a single particular<br />

habitat, mussel beds. Recently, based on presence/absence of species among all<br />

hydrothermal vent fields in the global ocean, Desbruyères et al. (2006a) and then Bachraty<br />

et al. (2009) argued that the Northern EPR (NEPR) and Southern EPR (SEPR) separate on<br />

both si<strong>de</strong>s of the Equator as two distinct provinces. Despite a more or less continuous<br />

distribution of the vent fauna along the SEPR, assemblages consist in a succession of small<br />

patches dominated either by vestimentiferan tubeworms, bivalves, or stalked cirripeds,<br />

each of them resembling assemblages typical of 13°N/9°50’N/EPR for tubeworms,<br />

21°N/EPR for bivalves, or the western Pacific back-arc basins for cirripeds (Jollivet et al.,<br />

2005). These observations led to the hypothesis that this ridge portion may represent a<br />

transition zone between the NEPR fauna and those of the Atlantic or the Western Pacific<br />

back arc basins where species may have reached their range limits (Jollivet et al., 2005).<br />

The BIOSPEEDO cruise, held in the south-east Pacific in 2004 (Jollivet et al.,<br />

2005), aimed at un<strong>de</strong>rstanding biodiversity patterns, and processes governing them, along<br />

262


Chapitre 5<br />

the SEPR: a poorly explored but important ridge cross-road between biogeographic vent<br />

provinces. The objective of the present study was twice:<br />

(1) to test the hypothesis that the SEPR could represent a transition zone between<br />

the NEPR and the other biogeographic provinces found in Atlantic and the<br />

Western Pacific. For this, macrofaunal community structure were compared at<br />

two spatial scales: (1) along the whole East Pacific Rise from 21°N to 17°S<br />

using species lists and then (2) along the 7°25’S-21°33’S/SEPR portion using a<br />

semi-quantitative dataset from 10 mussel beds at 6 hydrothermal vent fields in<br />

or<strong>de</strong>r to <strong>de</strong>termine the relative roles of local and regional factors on mussel bed<br />

diversities.<br />

(2) to generalize the hypothesis of vicariance across the Equator over nearly all<br />

vent habitats. This was assessed by a bar-co<strong>de</strong> analysis of the main gastropod<br />

species encountered along the EPR. Gastropods are in<strong>de</strong>ed an important<br />

component of hydrothermal fauna: species display various life history strategies<br />

and are found in all kinds of habitats (Jollivet, 1996; Mills et al., 2007).<br />

METHODS<br />

Taxa, regions and data sources<br />

Global analysis of communities along the EPR<br />

A global species list of benthic macrofauna including only the main fauna (i.e. gastropods,<br />

polychaetes and echino<strong>de</strong>rms) was generated from data compiled in (Desbruyères et al.,<br />

2006b), and completed with the species list of the BIOSPEEDO cruise. Highly mobile<br />

epifauna (e.g. crustaceans) have been intentionally exclu<strong>de</strong>d as they were not properly<br />

sampled during most hydrothermal cruises. A final list of species coming from chimney<br />

walls, vestimentiferan tubeworm clusters and mussel beds was thus compiled for nine<br />

hydrothermal vent fields: 21°N, 13°N, 9°50N, 7°25S, 14°S, 17°25S, 18°36S and 21°33S.<br />

In or<strong>de</strong>r to <strong>de</strong>tect more precisely biogeographic relationships, the data was analysed at the<br />

species presence-absence level (Desbruyères et al., 2006a, Bachraty et al. 2009).<br />

Variation of the macrofaunal community structure associated with mussel beds along the<br />

SEPR<br />

A total of nine semi-quantitative samples of Bathymodiolus thermophilus beds were<br />

collected in April/May 2004 along the Southern East-Pacific Rise (SEPR) at six vent fields<br />

from both volcanic-dominated and tectonic-dominated zones between 7°25’S and 21°33’S<br />

263


Chapitre 5<br />

(Table 5.1). Areas driven by volcanic activity consist in dome-shaped ridge crest where<br />

hydrothermal conduits resulting from seawater warming are short, shallower and unstable.<br />

Conversely, tectonic areas are characterized by a <strong>de</strong>ep axial graben where hydrothermal<br />

fluxes are controlled by faults and are more stable through time (Fouquet et al., 1994).<br />

Distances between neighbouring sites were comprised between 3 km and 891 km, and<br />

between 30 m and 147 m within a site (i.e. Grommit). Collections were performed using<br />

the hydraulic arm of the French manned-submersible vehicle ‘Nautile’ (Ifremer). For a<br />

brief <strong>de</strong>scription of all vent fields see Jollivet et al. (2005).<br />

264


Chapitre 5<br />

Table 5.1 Characteristics of the sampling units. Macrofauna <strong>de</strong>nsity corresponds to number of individuals standardize to the estimated<br />

surface sampled.<br />

265


Chapitre 5<br />

Mussels were washed on <strong>de</strong>ck with filtered seawater and the associated fauna was<br />

collected onto 1 mm mesh sieve. The animals were sorted onboard the ship; the retained<br />

organisms were fixed in 10% neutral formalin or in absolute ethanol for subsequent<br />

molecular analysis. In the laboratory, all specimens were sorted, i<strong>de</strong>ntified to the lowest<br />

possible taxon, i.e. mainly to the species level, un<strong>de</strong>r a dissecting microscope and then<br />

transferred to 70° ethanol. Data on echino<strong>de</strong>rm species correspond to those reported in<br />

Stöhr & Segonzac (2006). Large and highly motile predators, including zoarcid fishes,<br />

squat lobsters (Munidopsis spp) and bythograeid crabs were not inclu<strong>de</strong>d in the analysis.<br />

Their presence was assessed by screening dives’ vi<strong>de</strong>o records. Meiofauna and vagile<br />

epifauna (mainly amphipods and copepods), which were not a<strong>de</strong>quately sampled, were<br />

also exclu<strong>de</strong>d from the analyses. Mussels < 10 mm, usually <strong>de</strong>emed to represent an<br />

associated fauna at this life history stage rather than a structural component (Van Dover,<br />

2002; Dreyer et al., 2005), were also not inclu<strong>de</strong>d in this study. The proportion of mussel<br />

specimens < 10 mm, consi<strong>de</strong>red as juveniles, was however regar<strong>de</strong>d here as an indication<br />

of the mussel beds’ condition (newly settled, mature or senescent mussel beds).<br />

Surface area (SA) of mussel shells was calculated by applying the relationship set<br />

between the mass of aluminium foil required to cover the outer surface of a mussel shell<br />

and mussel length (Bergquist et al., 2005). One hundred and six mussel shells,<br />

representative of the whole size range, were used to <strong>de</strong>termine this relationship. The outer<br />

shells of each collection were wrapped in aluminium foil (0.00363 g.cm -2 ), and the mass<br />

required to cover the entire shell was then converted to surface area. Bias emerging from<br />

folding of the aluminium foil was consi<strong>de</strong>red equivalent for each mussel such that<br />

comparison between sampling units was valid. The relationship between SA (in cm 2 ) and<br />

shell mussel length (L) was <strong>de</strong>rived by plotting the surface area against the shell length to<br />

obtain the following equation:<br />

SA = 0.0137 L 1.9671<br />

This highly significant relationship (R² = 0.996; p =


Chapitre 5<br />

the shell, homogenized in 600µL of a 60°C preheated 2% CTAB buffer solution<br />

(containing 1.4M NaCl, 0.2% 2-mercaptoethanol, 20 mM EDTA, 100 mM Tris-HCl pH 8,<br />

0.1 mg/mL proteinase K) and digested for 2 h at 60°C. Proteins, lipids and carbohydrates<br />

were removed using a single chloroform-isoamyl alcohol (24:1) precipitation. DNA was<br />

precipitated with 370µL cold 100% isopropanol and stored at −20°C for 1 to 2 h. The tubes<br />

were finally centrifuged at 15 000g for 5 min. DNA pellets were washed with 70–75%<br />

ethanol and resuspen<strong>de</strong>d in 20–60µL of TE buffer (10 mM Tris-HCl; 20 mM EDTA pH 8).<br />

For each individual, partial sequences of mitochondrial cytochrome oxidase subunit<br />

I gene (mtCOI) were amplified by conducting a “nested PCR”. DNA was pre-amplified<br />

with the universal primers M1 and M2 (Nelson & Fisher 2000) and used as a target (1 µl of<br />

PCR-product) for a second amplification with the universal primers LCO1490 and<br />

HCO2198 <strong>de</strong>scribed by Folmer et al. (1994). For both amplifications, the 50 µl<br />

amplification mixture contained 3 µl of template DNA, 1x PCR buffer, 2 mM MgCl2, 0.4<br />

µM of each primer, 2.5 µM of each dNTP, 2 U Taq DNA polymerase, and sterile H2O to<br />

final volume. Polymerase chain reactions (PCR) were performed as follows: (a) a 3 min<br />

initial <strong>de</strong>naturation step at 94°C, (b) 40 cycles of 45 s of <strong>de</strong>naturation at 94°C, 45 s of<br />

annealing at 50°C and 90 s of elongation at 72°C and, (c) a 7 min final elongation at 72°C.<br />

The PCR products were purified and sequenced on ABI 3100 using BigDye®<br />

terminator chemistry (Applied Biosystems) following the manufacturer’s protocol.<br />

Sequences were proofread and aligned manually using BioEdit Sequence Alignment 7.0.1.<br />

Phylogenetic relationships between specimens were assessed for the wi<strong>de</strong>-range<br />

species found at a high number in the sample: the Lepetodrilids, i.e. Lepetodrilus elevatus,<br />

L. cristatus, L. pustulosus, L. tevnianus, Gorgoleptis emarginatus and G. spiralis, the<br />

Neomphalids, i.e. Pachy<strong>de</strong>rmia laevis and Peltospira operculata and the Neolepetopsid<br />

Eulepetopsis vitrea. Additional species, found in low proportion during the Biospeedo<br />

cruise but well known from the East Pacific Rise were also ad<strong>de</strong>d, i.e. Nodopelta heminoda<br />

and Rhyncopelta concentrica (Neomphalids), Clypeosectus <strong>de</strong>lectus and Gorgoleptis<br />

spiralis (Lepetodrilids). Details of specimens analysed are presented in Table 5.2. Aligned<br />

sequences were analysed by Bayesian inferences. First, the optimal substitution mo<strong>de</strong>l for<br />

COI was selected using the hierarchical likelihood ratio test (hLRTs) implemented in<br />

Mo<strong>de</strong>ltest 3.07 (Posada and Crandall 1998). Then, bayesian inferences were performed<br />

using the MrBayes 3.1.1 program (Huelsenbeck & Ronquist 2003) that uses a Markov<br />

Chain Monte Carlo (MCMC) method for exploring the parameter space in a stepwise<br />

267


Chapitre 5<br />

fashion. The analysis, involving 6 chains, was run for 10,000,000 generations with a<br />

sampling frequency of 1,000 and a burn_in period of 50,000. The trees were unrooted.<br />

Table 5.2 Number of specimens per species and per locality used in the bar-coding<br />

analysis<br />

Species This study Genebank Accession nos<br />

Eulepetopsis vitrea 21°N (1), 13°N (1), 9°50’N (1),<br />

7°25’S (1), 14°S (1), 17°25’S (1),<br />

17°34’ (1), 18°36’(1), 21°33’S (1)<br />

Lepetodrilus elevatus 21°N (1), 13°N (2), 14°S (1),<br />

17°25’S (1), 17°35’S (1), 21°33’S<br />

(4)<br />

Lepetodrilus aff. elevatus<br />

Lepetodrilus aff. galriftensis<br />

Lepetodrilus ovalis 7°25’S (1), 14°S (1), 17°25’S (1),<br />

17°35’S (1), 18°34’S (1), 21°33’S<br />

(1)<br />

EU306401-402 (Wood :<br />

22°N/EPR :21°N-0°)<br />

EU306407-408 (EPR:7°S-38°S)<br />

EU306413-414 (EPR:9°N-23°S)<br />

EU306478-479 (EPR:Monterey<br />

Bay/EPR:21°N-38°S)<br />

Lepetodrilus pustulosus 13°N (3) EU306457-458<br />

(Wood:22°N/EPR:21°N-17°S)<br />

Lepetodrilus aff. pustulosus<br />

EU306464-465 (EPR:31°S-38°S)<br />

Lepetodrilus tevnianus<br />

EU306389-390 (EPR:9°N-23°S)<br />

Lepetodrilus aff. tevnianus<br />

EU306395-396 (EPR:23°S-31°S)<br />

Lepetodrilus cristatus 13°N (17) EU306425-426 (EPR:21°N-38°S)<br />

Gorgoleptis emarginatus 7°25’S (2), 14°S (1), 17°25’S (2),<br />

21° 33’S (1)<br />

Gorgoleptis spiralis 14°S (1), 17°25’S (2), 21° 33’S (3)<br />

Clypeosectus <strong>de</strong>lectus 17°25’S (2), 21°S (1), 23°S (2)<br />

Pachy<strong>de</strong>rmia laevis 21°N (3), GoC (1), 13°N (1),<br />

17°25’S (2), 17°34’S (1), 18°36’S<br />

(2), 21°33’S (2), 31°S (1)<br />

Peltospira operculata 13°N (4), 14°S (2), 17°25’S (1),<br />

17°34’S (1), 18°25’S (2), 21°33’S<br />

(2)<br />

Peltospira <strong>de</strong>licata<br />

Nodopelta subnoda 13°N (2)<br />

Rhyncopelta concentrica 13°N (2)<br />

AY923931<br />

Published mtCOI sequences from Lepetodrilus elevatus (Genbank Accession nos<br />

EU306401-EU306402 and EU306407-EI306408 <strong>de</strong>scribed as L. aff. elevatus (Johnson et<br />

al., 2008)), L. pustulosus (Genbank Accession nos EU306457-EU306458 and EU306464-<br />

EU306465 for L. aff. pustulosus (Johnson et al., 2008)), L. cristatus (Genbank Accession<br />

nos EU306425-EU306426), L. ovalis (Genbank Accession nos EU306478-EU306479), L.<br />

aff. galriftensis (Genbank Accession nos EU306413-EU306414) and L. tevnianus<br />

(Genbank Accession nos EU306389-EU306390 and EU306395-EU306396 for L. aff.<br />

tevnianus (Johnson et al., 2008)) were inclu<strong>de</strong>d. A published sequence of Peltospira<br />

<strong>de</strong>licata (Genbank accession number AY923931) was also inclu<strong>de</strong>d.<br />

268


Chapitre 5<br />

Biodiversity measurements<br />

For both global and regional analyses, the species richness (S), the average taxonomic<br />

distinctness (Δ + ), variation in taxonomic distinctness (Λ + ), and the Jaccard coefficient were<br />

calculated from species presence/absence matrice using the DIVERSE subroutine in<br />

PRIMER v5 (Clarke & Warwick, 2001).<br />

The average taxonomic distinctness (Δ + ) is the average taxonomic distance between<br />

all pairs of species (i and j) in a sampling unit by gracing these distances through a<br />

taxonomic tree such as (Warwick & Clarke, 2001):<br />

[]/(1)/2ijijss!+


Chapitre 5<br />

Van Dover (2002, 2003). In the context of this study, the main advantage of these two<br />

indices is that they are in<strong>de</strong>pen<strong>de</strong>nt of sample size or sampling effort.<br />

Pielou’s eveness (J’), Shannon-Weiner diversity in<strong>de</strong>x (H’) and rarefaction in<strong>de</strong>x<br />

ES (n) were calculated for each mussel bed sampling unit from species relative abundance<br />

matrices using the DIVERSE subroutine in PRIMER v5. The E(Sn) rarefaction in<strong>de</strong>x is the<br />

Expected Number of Species which estimates the number of species (S) in samples<br />

normalized (rarefied) to successively smaller sizes (Hurlbert, 1971). It was originally<br />

<strong>de</strong>signed to allow comparisons of species diversity from sampling units of very different<br />

size. This method assumes that samples are homogeneous with similar distribution of<br />

individuals among species and that species are randomly distributed. These assumptions<br />

are rarely observed in practice and rarefaction in<strong>de</strong>x will strongly be biased for small<br />

samples. In our case, the sampling units were normalized to the lowest common number of<br />

individuals found at Grommit (21°33°S), i.e. 600 (Table 5.1).<br />

Multivariate statistical analyses<br />

Prior to analyses, a Hellinger transformation was applied to the response dataset in or<strong>de</strong>r to<br />

preserve the Hellinger distance, which is more appropriate for community composition<br />

data, instead of the Eucli<strong>de</strong>an distance (Legendre & Gallagher, 2001).<br />

Global analysis of communities along the EPR<br />

For global analysis, putative biogeographic breaks were i<strong>de</strong>ntified using a multivariate<br />

regression tree (MRT) (De'ath, 2002) applied on two different datasets. The first one came<br />

from the species list recor<strong>de</strong>d in Desbruyères et al. (2006b). The species collected during<br />

the BIOSPEEDO cruise were ad<strong>de</strong>d to the second analysis. This method was used to<br />

<strong>de</strong>fine group of fields that are similar in faunal composition while geographically adjacent.<br />

The MRT is a partitioning method of the multivariate data response table (i.e. species<br />

matrix) constrained along the axis of a single explanatory variable. Latitu<strong>de</strong> was used as<br />

the constraining variable to generate spatially consistent reliable groups. The tree structure<br />

is generated by partitioning the whole dataset into mutually exclusive groups and is grown<br />

by successive binary splits of the data, each split involving a breakpoint in the explanatory<br />

variable. Each split is chosen to minimize the total multivariate sum of squares within the<br />

two groups formed, or maximise the R² (or equivalently the among-group sum of squares)<br />

at each division step. The tree with the lowest cross-validation error was chosen as the best<br />

predictive tree. At the end of the procedure, each leaf of the tree is characterized by the<br />

270


Chapitre 5<br />

number of sites, the multivariate mean of the sites and the explanatory variable values that<br />

<strong>de</strong>fine it. The tree was computed using the ‘mvpart’ library (De'ath, 2002) in the R<br />

statistical language (R Development CoreTeam, 2006).<br />

Variation of the macrofaunal community structure associated with mussel beds along the<br />

SEPR<br />

To investigate the correlations of faunal composition with explanatory variables, a<br />

redundancy canonical analysis (RDA) was performed on standardized data on macrofauna<br />

associated with mussel beds along the Southern EPR. This method combines aspects of<br />

ordination and regression and uses a permutation procedure to test the significance of the<br />

explained variation; thus species abundances have not to be normally distributed (Legendre<br />

& Legendre, 1998). The explanatory variables were the latitu<strong>de</strong> from the equator, the<br />

geodynamic environment (two-state variable: tectonic vs volcanic), the type of venting<br />

(three-state variable: moribund, diffusion zone or chimney wall) and the proportion of<br />

mussels juveniles (i.e.


Chapitre 5<br />

21°33’S. Similarly, Echinopelta fistulosa and Cyathermia naticoi<strong>de</strong>s, only known from the<br />

Northern EPR (i.e. 9°N to 21°N), were collected at 21°33’S and 17°25’S respectively. For<br />

the first time, benthic individuals of the genus Laeviphitus were collected while this taxon<br />

had only been found as larvae in this area (Desbruyères et al., 2006b). Furthermore, this<br />

cruise led to the discovery of new species, i.e. Phymorhyncus n. sp. at 17°25’S and<br />

18°25’S, Melanodrymia n. sp. and Haloceras sp., a new genus for hydrothermal vents, at<br />

21°33’S for gastropods (Warén, pers. obs.) and Paralvinella n. sp., Glycera n. sp. and a<br />

new Ampharetidae species at 21°33’S, Sphaerodopsis n. sp. at 18°25’S and Nicomache<br />

n.sp. and a new Flabelligeridae species at 17°25’S and 21°33’S for polychaetes<br />

(Desbruyères & Hour<strong>de</strong>z, pers. obs.), as well as species belonging to genera characterizing<br />

other ridges systems (i.e. Eosipho auzendi, Pacific Antarctic Ridge and Shinkailepas sp.<br />

from the Mid-Atlantic Ridge and back-arc basins). These new species were only found<br />

south of the latitu<strong>de</strong> of 17°25’S. Moreover, the gastropods Nodopelta heminoda and N.<br />

subnoda and the polychaete Hesiolyra bergi, highly abundant on the chimney walls of the<br />

NEPR vent fields, were virtually nearly absent (very few individuals) from SEPR samples<br />

<strong>de</strong>spite a huge sampling effort (9 chimneys extensively sampled).<br />

When analysing only the species dataset reported in the handbook, species richness<br />

was twice higher at all NEPR vent fields (i.e. 42 to 62 species) than at any vent field from<br />

SEPR (i.e. 13 to 28 species) (Table 5.3). The highest species richness was observed at<br />

13°N and 21°N (i.e. 62 and 51, respectively) and at 17°25’S (i.e. 28) for the southern part.<br />

Conversely average taxonomic distinctness was nearly homogeneous along the whole EPR<br />

with the highest value (84.48) found at 17°25’S. Adding the species recor<strong>de</strong>d during the<br />

BIOSPEEDO cruise increased significantly the number of species recor<strong>de</strong>d in the SEPR to<br />

an extent similar to that of the NEPR. The analysis revealed two centres of species<br />

diversity: one in the north including 11°N and 13°N and one in the south inclu<strong>de</strong>d vent<br />

fields located between 17°25’S and 21°33’S. Species diversity in the SEPR reached 60 at<br />

17°25’S against 62 at 13°N. The average taxonomic distinctness was not too much affected<br />

with a highest value still found along the SEPR and associated with the highest variation<br />

coefficients of the taxonomic distinctness (Table 5.3).<br />

272


Chapitre 5<br />

Table 5.3 Diversity indices for each hydrothermal vent field along the East-Pacific<br />

Rise (EPR).<br />

Data are expressed in presence/absence. EPR handbook: data reported in (Desbruyères et<br />

al., 2006b) and EPR hb+bs: species collected in the Southern EPR during the<br />

BIOSPEEDO cruise were ad<strong>de</strong>d to the data.<br />

EPR hand book<br />

EPR hb + bs<br />

Vent site S Δ + Λ + S Δ + Λ +<br />

21°N 51 79.79 513.2 51 79.79 513.2<br />

13°N 62 80.94 536.2 62 80.94 536.2<br />

11°N 42 80.22 560.4 42 80.22 560.4<br />

9°50’N 45 79.41 566.3 45 79.41 566.3<br />

7°25S 21 78.73 555.5 35 79.33 519.4<br />

14°S 22 79.73 520.9 34 79.32 557.8<br />

17°S 28 84.48 445.5 60 81.96 484.9<br />

18°S 14 79.3 472.2 38 83.95 398.7<br />

21°S 13 76.5 473.2 52 80.62 507.1<br />

The multiple cross-validations produced a MRT with three groups of fauna for both<br />

analyses, explaining 70.3% of the variation in the species dataset when consi<strong>de</strong>ring only<br />

the handbook and 45.4% when adding the new species records (Fig. 5.2). Those trees<br />

displayed the lowest cross-validation error and offered sufficient resolution when<br />

consi<strong>de</strong>ring the number of observed hydrothermal fields along the East Pacific Rise (i.e.<br />

9). The first split separated the northern and southern parts of the EPR more or less at the<br />

Equator. A second split sub-divi<strong>de</strong>d the SEPR into two groups of assemblages including<br />

sites on both si<strong>de</strong>s of the latitu<strong>de</strong> of 18°S when consi<strong>de</strong>ring only species reported in the<br />

Handbook, and between 14°S and 17°25’S when adding species collected during the<br />

BIOSPEEDO cruise.<br />

273


Chapitre 5<br />

Fig. 5.2 Groups obtained by multivariate regression trees (MRT) analysis of<br />

presence/absence data along the EPR. Latitu<strong>de</strong>s are south from the equator. A. Data<br />

from the handbook (Desbruyères et al., 2006b). B. Data complemented with species data<br />

from the BIOSPEEDO cruise.<br />

274


Chapitre 5<br />

Evi<strong>de</strong>nce of species crypticism and vicariance in vent gastropods<br />

Phylogenetic relationships were analysed for 18 species of Neolepetopsidae,<br />

Lepetodrilidae and Neomphalidae gastropods along the whole EPR using a 425pb<br />

sequence portion of the Cytochrome oxidase subunit I (COI). The tree obtained by<br />

Bayesian inferences is presented in Fig. 5.3. All no<strong>de</strong>s with less than a 95% probability<br />

were collapsed to basal polytomy. Within the Lepetodrilidae family, we <strong>de</strong>tected thirteen<br />

cla<strong>de</strong>s with ten cla<strong>de</strong>s involving six species of the genus Lepetodrilus sp.. The two species<br />

of the genus Gorgoleptis sp. and Clypeosectus <strong>de</strong>lectus were constituted of one cla<strong>de</strong> each.<br />

Pairwise K2P Kimura distances between cla<strong>de</strong>s <strong>de</strong>tected within each well-i<strong>de</strong>ntified<br />

morphological species, i.e. L. elevatus, L. pustulosus, and L. tevnianus, ranged between<br />

5.5% and 8.3%. Within cryptic lineages (i.e. cla<strong>de</strong>s), pairwise distances ranged from 0 to<br />

1.5% for L. ovalis. Lepetodrilus aff galriftensis, i<strong>de</strong>ntified by Johnson et al. (2008),<br />

grouped with the southern cla<strong>de</strong> (9°50N to 21°33S) of Lepetodrilus elevatus (Matabos et<br />

al. 2008, Plouviez et al. 2009) and Lepetodrilus aff elevatus (Johnson et al. 2008) with the<br />

third cla<strong>de</strong> only found at 21°33S. This latter diverged from the northern and southern<br />

cla<strong>de</strong>s by 7% and 5.8% respectively. L. ovalis and G. emarginatus also displayed two<br />

cla<strong>de</strong>s with a slight divergence of 1.5%. G. spiralis and C. <strong>de</strong>lectus were homogeneous<br />

along the EPR with intra-species distance comprised between 0.01 and 1.5%. Eulepetopsis<br />

vitrea displayed two distinct mitochondrial lineages separated by a divergence of 1.5% and<br />

located on each part of the Equator. The only gastropod species of the genus Lepetodrilus<br />

for which no geographic break was <strong>de</strong>tected is L. cristatus, mostly found on the walls of<br />

vent chimneys. Within the Neomphalinae, Peltospira operculata and Pachy<strong>de</strong>rmia laevis<br />

exhibited two to three divergent cla<strong>de</strong>s. Peltospira operculata and Pachy<strong>de</strong>rmia laevis<br />

showed the strongest divergence between the NEPR and SEPR with 7.9% and 12.5%<br />

divergence between the two cla<strong>de</strong>s respectively. An individual of P. laevis sequenced from<br />

31°S beyond the Easter microplate showed 10.8% divergence with the NEPR cla<strong>de</strong> and<br />

11.2% with the SEPR, respectively. Pairwise K2P Kimura distances within cla<strong>de</strong>s ranged<br />

from 0.2% to 0.6%. In all cases, cryptic lineages within a single morphological species<br />

were geographically structured with a northern and southern lineage across the Equator<br />

and, in a few cases an additional mt lineage at the most southern vent field (i.e. 21°33S).<br />

However the absence of samples between 9°50’N and 17°25’S impe<strong>de</strong> to clearly i<strong>de</strong>ntify<br />

the localisation of the geographic break along the ridge, and as it was impossible to<br />

i<strong>de</strong>ntify lineages on the basis of morphological criteria, the species complexes i<strong>de</strong>ntified<br />

275


Chapitre 5<br />

here were consi<strong>de</strong>red as a single wi<strong>de</strong>ly distributed species in further biogeographical<br />

analyses.<br />

276


Chapitre 5<br />

Fig. 5.3 Bayesan tree based on the 111 gastropods COI sequences. Numbers correspond<br />

to the posterior probability values of the no<strong>de</strong>s. Genebank sequences ad<strong>de</strong>d to the tree were<br />

provi<strong>de</strong>d by (Johnson et al., 2008). Abbreviations for sites are given in Table 3.<br />

277


Chapitre 5<br />

Variations in the macrofaunal community structure of mussel beds along the SEPR<br />

Up to 13,339 individuals belonging to sixtie species were collected in mussel beds along<br />

the Southern EPR: twenty-height species of gastropods (nineteen genera and eleven<br />

families), thirty species of polychaetes (twenty-two genera and fourteen families) and 2<br />

species of echino<strong>de</strong>rms (two genera and two families) (Table 5.4). The most abundant<br />

species at the regional scale were Lepetodrilus sp. (i.e. L. elevatus and L. ovalis),<br />

Eulepetopsis vitrea, Amphisamytha galapagensis and Archinome rosacea. While<br />

Melanodrymia aurantiaca was only found at 21°33’S, it was highly abundant at that site.<br />

Conversely, Planorbi<strong>de</strong>lla sp. was found in all samples but represented few individuals at<br />

each locality.<br />

278


Chapitre 5<br />

Table 5.4 List of species collected in 10 mussel beds along the Southern East-Pacific Rise (SEPR).<br />

279


Chapitre 5<br />

Table 5.4 (continued…)<br />

280


Chapitre 5<br />

Local species richness varied between 11 and 26 with the lowest values observed at Lucky<br />

Eric (14°S), Susie (17°34’S) and Sarah’s Spring (7°25’S) (Table 5.5). The highest species<br />

diversities, both in terms of species richness and evenness, were observed at the 17°25’S<br />

Rehu Marka and 18°36’S Animal Farm sites and at one sample from the 21°33’S Grommit<br />

site. However, Shannon indices are relatively low, ranging between 0.47 at Suzie and 2.33<br />

at Rehu Marka (Table 5.5).<br />

Table 5.5 Diversity indices for each quantitative sample collected in mussel beds<br />

along the Southern East-Pacific Rise (SEPR).<br />

Vent field Vent site N(ind.m²) S ES (600) J’ H’(log e ) Δ + Λ +<br />

7°24S Sarah’s Spring 3959 12 11.64 0.52 1.29 75.45 798<br />

14°S Lucky Eric 1170 11 10.65 0.51 1.22 67.20 980<br />

17°25S Oasis-BS6 2028 26 17.44 0.20 0.65 87.02 310<br />

Rehu Marka 771 24 22.48 0.75 2.33 86.01 374<br />

17°34S Susie 1794 12 8.85 0.19 0.47 77.09 835<br />

18°36S Animal Farm 2249 24 19.5 0.71 2.24 88.84 258<br />

21°33S Grommit (Gr1) 647 25 25.24 0.39 1.26 84.6 431<br />

Grommit (Gr2) 832 26 25.6 0.68 2.23 86.28 335<br />

Grommit (Gr3) 845 24 22.2 0.55 1.74 84.2 472<br />

The same trend was observed regarding to average taxonomic distinctness indices.<br />

Δ + varied between 67.20 for Lucky Eric (14°S) and 88.84 for Animal Farm (18°36’S). The<br />

following highest values, observed for collections at Oasis (17°25’S), Rehu Marka<br />

(17°25’S) and Grommit (21°33’S), were equivalent. Taxonomic groups were more evenly<br />

distributed at sites characterized by the highest average taxonomic distinctness, as<br />

suggested by the Λ + value. Oasis (17°25’S), Rehu Marka (17°253S), Animal Farm<br />

(18°36’S) and one replicate from Grommit (i.e. Gr2, 21°33’S) were significantly higher<br />

than the theoretical value estimated by 1000 permutations of the dataset. The sampling unit<br />

collected at Lucky Eric (14°S) was significantly smaller (Fig. 5.4) in term of average<br />

taxonomic distinctness; and the variation in taxonomic distinctness values observed at<br />

Sarah’s Spring (7°25’S), Lucky Eric (14°S) and Susie (17°34’S) differed significantly<br />

from the theoretical value (Fig. 3).<br />

281


Chapitre 5<br />

Fig. 5.4 Measured values of Δ + (a) and Λ + (b) from the semi-quantitative sampling<br />

units collected in mussel beds along the SEPR, plotted against the number of species<br />

on the simulated 95% confi<strong>de</strong>nce interval generated after 1000 resamplings of the<br />

same number of species from the total SEPR species list. Sites abbreviations are<br />

presented in Table 1.<br />

282


Chapitre 5<br />

The cluster analysis shown in Fig. 5.5 indicated a clear separation of sampling sites into<br />

three distinct groups plus one isolated site, Sarah’s Spring (7°25’S). Sarah’s Spring was<br />

characterized by low species richness and taxonomic diversity, with the absence of<br />

echino<strong>de</strong>rms and the presence of only four species of polychaetes. This site differs from<br />

the others in that the gastropod Lepetodrilus pustulosus, almost absent from other samples,<br />

was found in a high proportion. L. pustulosus and the polychaete Amphisamytha<br />

galapagensis constituted 85% of the individuals collected. The first group corresponds to<br />

the highest values of the Shannon’s diversity in<strong>de</strong>x (i.e. H’) and a high average taxonomic<br />

distinctness (Table 5.5) and inclu<strong>de</strong>s Animal Farm (18°36’S) and Rehu Marka (17°25’S).<br />

They were both composed of five and seven dominant species (i.e. >5%). The second<br />

group, which inclu<strong>de</strong>s collections from Lucky Eric (14°S), Oasis BS6 (17°25’S) and Susie<br />

(17°34’S), corresponds to the lowest Shannon’s diversity in<strong>de</strong>x (Table 5.5). They were<br />

mainly composed of Lepetodrilid species, with L. elevatus present up to 88% and 90% in<br />

the sampling units collected at Oasis BS6 and Susie, respectively. Finally, samples<br />

collected at Grommit (21°33’S) grouped together with 65% of species similarity. They<br />

displayed the same species and taxonomic diversity and contained at least three new<br />

species each, and six new species only shared by those samples (i.e. Melanodrymia n. sp.,<br />

Shinkailepas sp., Glycera n. sp., Nichomache n. sp., Oasisia n. sp. and the new species of<br />

Ampharetidae). While the gastropod Melanodrymia auriantaca was completely absent<br />

from the other sampling units, it was found at a very high abundance at this site where<br />

samples were collected in mussels beds from the chimney wall.<br />

Fig. 5.5 Group-average sorting <strong>de</strong>ndrogram computed from the cord distance among<br />

the 9 sampling units collected in mussel beds along the Southern East Pacific Rise.<br />

Numbers i<strong>de</strong>ntify clusters highlighted. Sites abbreviations are presented in Table 1.<br />

283


Chapitre 5<br />

Sites and explanatory variables used in the RDA are displayed as an ordination<br />

biplot in Fig. 5.6. Results of the RDA showed that the transformed species abundances are<br />

significantly correlated with two environmental variables out of four (R²adj=0.29,<br />

p=0.004). The first axis explained 28.6% (R²adj=0.18, p=0.003) of the total variation in<br />

species abundance data, and the latitu<strong>de</strong> and the type of venting explained 27.8%<br />

(R²adj=0.17) and 26.4% (R²adj=0.16), respectively. However, the power of the test is very<br />

low due to the lack of observations and environmental variables in the dataset. This<br />

prevents the analysis from reaching a statistical significance when consi<strong>de</strong>ring each<br />

variable alone.<br />

Fig. 5.6 Ordination correlation biplot of the sites and environment variables based on<br />

redundancy analysis (RDA) of transformed species abundance data. Abbreviations used<br />

for the sites are presented in Table 1. Groups i<strong>de</strong>ntified with the <strong>de</strong>ndrogram are represented<br />

with circles. Species were not represented.<br />

284


Chapitre 5<br />

DISCUSSION<br />

Evi<strong>de</strong>nce of crypticism in gastropod species<br />

For the Lepetodrilids species, 6.5% divergence was already highlighted for Lepetodrilus<br />

elevatus separating the northern and southern cla<strong>de</strong>s found in sympatry at 9°50N (Matabos<br />

et al., 2008b). Johnson et al. (2008) also revealed several distinct cryptic species<br />

complexes within the genus Lepetodrilus with pairwise distances ranging between 4.37<br />

and 31.25%. Pachy<strong>de</strong>rmia laevis displayed 12.5% divergence between north and Southern<br />

EPR, higher than the divergence found with the individual collected on the other si<strong>de</strong> of<br />

the microplate. Differentiation between individuals of Planorbi<strong>de</strong>lla planispira found on<br />

both si<strong>de</strong>s of the equator fell in the same range as the Lepetodrilids species with 7.9%<br />

divergence. Strong genetic differentiation between Northern and Southern EPR were<br />

<strong>de</strong>tected for other divergent taxonomic groups including bivalves and polychaetes<br />

(Plouviez et al., 2009). While levels of divergence can be highly variable between two<br />

species (Knowlton, 2000), pairwise genetic distances between and among cla<strong>de</strong>s<br />

highlighted in this study fit inter- and intra-specific levels respectively (Peek et al., 1997).<br />

Since the <strong>de</strong>velopment of molecular techniques, occurrence of sibling species appears to be<br />

more wi<strong>de</strong>spread in hydrothermal ecosystems, and at larger scope, in the sea (Peek et al.,<br />

1997; Knowlton, 2000; Johnson et al., 2008; Matabos et al., 2008b).<br />

Biogeographic entities along the EPR<br />

Barcoding of hydrothermal species (Johnson et al., 2008), population genetics studies (e.g.<br />

Chevaldonné et al., 2002; Won et al., 2003; Hurtado et al., 2004; Matabos et al., 2008b)<br />

and comparative phylogeography (Plouviez et al., 2009) along the EPR converge to the<br />

existence of a population genetic break on both si<strong>de</strong>s of the equator. Presence of<br />

topographic barriers in relation with tectonic events (i.e. formation of transform faults,<br />

microplates, bathymetric inflation) favoured the role of vicariance events and subsequent<br />

population differentiation at hydrothermal vents (Tunnicliffe & Fowler, 1996). Several<br />

topographic features, related to the ridge history, have been proposed to explain the<br />

divergences observed within morphological species. The formation of the Bauer<br />

microplate, 17-6 Mya (Matabos et al., 2008b), and the nearly simultaneous fossilization of<br />

the Mathematician ridge (Plouviez et al., 2009) were proposed to account for the<br />

divergence observed in the split of Lepetodrilus elevatus into two cryptic lineages. Four<br />

out of six Lepetodrilid species display the same range of genetic divergence (i.e. 5.2 to 9.2;<br />

285


Chapitre 5<br />

Johnson et al., 2008; Matabos et al, 2008b; this study) and may thus share the same past<br />

history. However, levels of divergence are variable among species with two sets of values<br />

ranged from 6-12% (an ‘old’ vicariant event involving L. elevatus, L. pustulosus, L.<br />

tevniatus, P. operculata, P. laevis) to 1.5-3% (‘recent’effect involving L. ovalis, E. vitrea,<br />

G. emarginatus), respectively. Tests on shared divergence times were performed on seven<br />

hydrothermal vent species along the EPR by Plouviez et al. (2009) and revealed the<br />

occurrence of a common vicariant event ~1.3 Mya for all studied species, including L.<br />

elevatus for which another ol<strong>de</strong>r speciation event is also <strong>de</strong>tected around 11.5 Mya. These<br />

latter authors therefore proposed that the formation of transform faults around 1 to 2 Mya<br />

between 4°S and 14°S, may be responsible for this vicariant break and the emergence of a<br />

transition zone in this area.<br />

Analyses of community composition expressed in terms of presence/absence<br />

support the molecular dataset. Recently, Bachraty et al. (2009) suggested that the EPR is<br />

separated into two different biogeographic provinces: the NEPR vs the SEPR. Our MRT<br />

analyses confirmed the presence of a geographic break at the equator separating the EPR<br />

fauna into two biogeographic entities. However, species lists and molecular data between<br />

9°N and 7°25’S are lacking and further studies centered on the Hess Deep are nee<strong>de</strong>d to<br />

exactly position the break. The two evolutionary lineages of Lepetodrilus elevatus were<br />

found in sympatry at 9°50N (Matabos et al., 2008b) were and able to hybridize so other<br />

cryptic complex species may also overlap around the equator. This is especially the case<br />

for Lepetodrilus ovalis and Gorgoleptis emarginatus which showed a consistent but weak<br />

divergence between two distinct mitochondrial lineages that overlap over nearly all the<br />

latitudinal gradient, with one lineage dominating the northern part of the EPR. A contrario,<br />

the gastropod E. vitrea, which displays the same level of divergence exhibits a complete<br />

geographic separation of its two cryptic lineages at 7°25’S (Plouviez et al. 2009). One<br />

explanation could be the differences in life history strategies that can influence dispersal<br />

abilities. Some species may have in<strong>de</strong>ed a dispersal potential to cross the barrier and meet<br />

secondarily. Genome homogenisation between the previously isolated cla<strong>de</strong>s is then<br />

<strong>de</strong>pending on the strength of selection against hybrids (genetic barrier).<br />

Ad<strong>de</strong>d to the occurrence of cryptic species from both parts of the EPR, the<br />

BIOSPEEDO cruise brought at least ten new species for hydrothermal vents (three<br />

gastropods, six polychaetes and one echino<strong>de</strong>rm; Stöhr & Segonzac, 2006) and expands<br />

the range of several species previously <strong>de</strong>scribed from the NEPR. The wi<strong>de</strong> geographic<br />

range of most vent species confirms studies previously conducted by Van Dover (2002) in<br />

286


Chapitre 5<br />

the same area. However, while the first studies based on megafauna observations<br />

(Geistdoerfer et al., 1995; Halanych et al., 1999) and the quantitative analysis of<br />

communities associated with mussel beds (Van Dover, 2002, 2003) conclu<strong>de</strong>d that the<br />

fauna was similar between the Northern and Southern EPR, results from the present study<br />

highlighted several latitudinal changes in community composition between 17°S and 21°S.<br />

New species were found beyond 17°S and some species warrant further studies as their<br />

i<strong>de</strong>ntifications were uncertain (Hour<strong>de</strong>z & Desbuyères; A. Warén, pers com.). In the listed<br />

taxa, two species, Shinkailepas sp. and Eosipho auzendi, represented by two and more<br />

individuals respectively, belong to a genus so far known only from the Mid-Atlantic Ridge<br />

and the Northwest Pacific back-arc basins for Shinkailepas sp. (Desbruyères et al., 2006b).<br />

The presence of Eosipho was already reported at Rehu Marka by (Van Dover, 2002).<br />

Ad<strong>de</strong>d to the presence of at least five cryptic species (i.e. Lepetodrilus elevatus, L.<br />

pustulosus, L. tevnianus, Planorbi<strong>de</strong>lla planispira, Pachy<strong>de</strong>rmia laevis), the East-Pacific<br />

Rise appears to be a more complex biogeographic zone than it was first suggested. When<br />

taking into account only the species recor<strong>de</strong>d in Desbruyères et al. (2006b), a geographic<br />

break was established at 18°S and the highest species diversity was found at 13°N. When<br />

adding species collected on the SEPR during the BIOSPEEDO cruise, species diversity<br />

was not different anymore between NEPR and SEPR vent sites and the geographic break<br />

along the SEPR appears positioned between 14°S and 17°25’S. The hypothesis of<br />

Bachraty et al. (2009) that the NEPR could be a centre of dispersal, is probably due to the<br />

over-sampling processed on the northern part of the EPR. Observations and molecular data<br />

led to the assumption that the high species diversity observed on this part of the SEPR is<br />

the result of the overlap of several distinct biogeographic provinces. The new species of<br />

Phymorhyncus sp. had been already found at 31°S (A. Warén, pers. obs.) in concordance<br />

with the presence of the gastropod Eosipho auzendi, <strong>de</strong>scribed from the Pacific Antarctic<br />

ridge. Individuals of Shinkailepas sp., a genus only <strong>de</strong>scribed on the Mid-Atlantic Ridge<br />

and in the back-arc basins, were encountered at 21°33’S. Furthermore the average<br />

taxonomic distinctness for the sampling units collected in mussel beds located between<br />

17°25’S and 21°33’S (i.e. Oasis BS6, Rehu Marka, Animal Farm and Grommit) <strong>de</strong>viated<br />

significantly from the theoretical number of expected species, reinforcing the hypothesis of<br />

“abnormal” hot spot of diversity in this area. The southern part of the SEPR (i.e. between<br />

17°25’S and 21°33’S) could be thus a transition (mixing) zone between the NEPR, the<br />

Pacific Antarctic ridge, and possibly the back-arc basins of the western Pacific, supporting<br />

first observations ma<strong>de</strong> during the BIOSPEEDO cruise (Jollivet et al., 2005). This<br />

287


Chapitre 5<br />

assumption is supported by the co-occurrence of NEPR-like Alvinella colonies and<br />

Chorocaris sp. swarms at 21°33’S, shrimp swarms being one of the most striking<br />

characteristic of the Mid-Atlantic Ridge, the Indian Ridge and the Northwest Pacific backarc<br />

basins (Van Dover et al., 2001a; Jollivet et al., 2005).<br />

Mussel beds community structure and composition along the Southern East-Pacific<br />

Rise<br />

The latitu<strong>de</strong> gradient in the global MRT analysis accounted for 45.4% of the variation in<br />

the presence/absence species dataset when adding species collected during the<br />

BIOSPEEDO cruise, suggesting additional role of regional and local factors on the<br />

community composition. Diversity associated with mussel beds was thus analysed along<br />

the SEPR to better un<strong>de</strong>rstand processes influencing community structure at the regional<br />

scale. Diversity associated with mussel beds has been extensively studied along the East-<br />

Pacific Rise (Van Dover & Trask, 2000; Van Dover, 2002; Turnipseed et al., 2003; Van<br />

Dover, 2003; Bergquist et al., 2005; Dreyer et al., 2005; Van Dover & Doerries, 2005).<br />

Most of those studies were conducted from a quantitative sampling including vagile<br />

epifauna and large macrofauna juveniles (e.g. crabs). This difference can account for the<br />

lower species richness observed in this study compared to previous studies performed<br />

along the SEPR (Van Dover, 2002, 2003). Whatever the sampling strategy used, diversity<br />

recor<strong>de</strong>d along the SEPR was always higher than values reported from the NEPR (Van<br />

Dover, 2003; Dreyer et al., 2005). Results from the canonical analysis showed that latitu<strong>de</strong><br />

combined with the type of venting at a regional scale significantly influenced the<br />

community structure and composition associated with mussel beds along the SEPR. The<br />

Oasis (17°25’S), Rehu Marka (17°24’S), Animal Farm (18°36’S) and Grommit (21°33’S)<br />

sites were characterized by a high diversity when compared to the most northern sites. At<br />

Sarah’s Spring (7°25’S) and Lucky Eric (14°S) sampling localities, mussels were<br />

distributed in a scattered fashion which could explain the low diversity observed. In<strong>de</strong>ed,<br />

the low diversity observed was related to a low number of polychaete species in the<br />

sampling units. This is further supported by the significant <strong>de</strong>viation of the variation in<br />

taxonomic distinctness from theoretical values for those sampling units.<br />

The significant influence of the type of venting highlighted the role of local<br />

processes on the patterns observed. Edifice life spans are highly variable within a single<br />

vent field in response to stochastic events altering fluid discharge and mineralisation<br />

processes. Variations in fluid discharge intensity greatly influence physico-chemical<br />

288


Chapitre 5<br />

parameters and biotic interactions between species (Mullineaux et al., 2003) which, will in<br />

turn greatly impact the community structure and composition. Species composition is thus<br />

highly variable in time and space within a single site and strong differences are observed<br />

according to the successional stage of the community at the time of sampling (Jollivet,<br />

1996; Sarrazin et al., 1997; Shank et al., 1998). On the EPR, productivity, species richness<br />

and abundances of invertebrates associated with mussel beds appeared to be lower at<br />

waning and ol<strong>de</strong>r sites than at young and active sites (Van Dover, 2003; Dreyer et al.,<br />

2005). In this study, mussel bed’s age was unknown, and the temporal state of mussel beds<br />

was assessed via <strong>de</strong>mographic structure of mussel populations and the proportion of<br />

juveniles. However mussel <strong>de</strong>mographic structure did not reflect in situ observations of the<br />

assemblages (i.e. presence of <strong>de</strong>caying vestimentiferan tubes at Sarah’s Spring) and no<br />

correlation has been found between community structure and the mussel size frequency<br />

distribution (unpublished data) or the proportion of mussels’ juveniles. This latter was high<br />

at the moribund site Animal Farm (i.e. 56%, Table 5.1) while a low proportion of juveniles<br />

characterized the large and <strong>de</strong>nse mussel beds <strong>de</strong>veloping in the waxing site Grommit<br />

(21°S) (i.e. between 0 and 9%). Thus proportion of mussel juveniles does not appear to be<br />

a good proxy to characterize mussel beds conditions.<br />

Analyses showed however that the type of venting had a significant influence when<br />

combined with the latitu<strong>de</strong>. First, the lack of sampling units along the gradient can impe<strong>de</strong><br />

the test to reach the level of significance when consi<strong>de</strong>ring the type of habitat alone. On the<br />

other hand, this result could reflect the intermingling role of local and regional factors.<br />

In<strong>de</strong>ed, processes acting at different spatial scales are often imbricate (Levin et al., 2001)<br />

such that it is difficult to i<strong>de</strong>ntify their relative role in the absence of replicate samples.<br />

Thus, although the influence of the geodynamic environment was not significant, the type<br />

of venting is strongly related to the regional dynamic of the ridge (Fouquet et al., 1994).<br />

All chimneys were found in vent fields characterized by a tectonic environment (i.e.<br />

Grommit-21°33’S and Susie-17°34’S). Areas driven by volcanic activity are more unstable<br />

in time and space than tectonic areas (Fouquet et al., 1994). Those differences, by<br />

controlling the extinction rate of sites, will impact the extinction/recolonisation rate of the<br />

vent fauna, promote habitat heterogeneity and reduce distances between sites (Jollivet et<br />

al. 1999). While unstable environments are expected to reduce species diversity (Juniper &<br />

Tunnicliffe, 1997), some authors suggested that the resulting small distances between vent<br />

sites should account for bursts of gene flow (Jollivet et al. 1999) and thus result in greater<br />

289


Chapitre 5<br />

rate of site recolonisation promoting a higher species diversity (Turnipseed et al., 2003;<br />

Van Dover & Doerries, 2005). In this study, species and taxonomic diversities were<br />

equivalent in the two types of vent dynamics, suggesting that differences resi<strong>de</strong>d in species<br />

composition. As an example, three gastropod species (i.e. Melanodrymia aurantiaca,<br />

Echinopelta fistulosa and Nodopelta subnoda), known to colonize active black smokers’<br />

chimney walls (Jollivet, 1996), were only collected at the Grommit site (21°33’S) where<br />

musselbeds <strong>de</strong>veloped at the base and on the wall of a black smoker. Similarly, scavenger<br />

species (i.e. Phymorhyncus spp., Eosipho auzendi, Spioniphura jolliveti) were mainly<br />

found in sites located in volcanic areas (i.e. Oasis BS6, Rehu Marka and Animal Farm).<br />

Nonetheless, while all sampling units collected in the tectonic 21°33’S vent field<br />

(i.e. Gommit) were similar in species composition, strong differences were observed at the<br />

volcanic 17°25’S area (i.e. Oasis BS6 and Rehu Marka). In unstable volcanic areas, the<br />

high rate of spatio-temporal changes in fluid fluxes will result in the intermingling of<br />

different successional stages within a site, thus enhancing the variance among mussel beds<br />

in a given area. Thus, while found on dome-shaped ridge, the apparent moribund sites<br />

Sarah’s Spring and Animal Farm displayed large differences in community structure. A<br />

low diversity characterized the sample collected at Sarah’s Spring with the overwhelming<br />

presence of the gastropod Lepetodrilus elevatus (48%) and the polychete Amphisamytha<br />

galapagensis (37%). Lepetodrilid species are grazers and are among the first species to<br />

colonize a new site, thus not representative of waning sites and an indicator of a vent reopening<br />

conditions. In<strong>de</strong>ed, the presence on this site of small specimens of the tubeworms<br />

Tevnia jerichonana and Oasisia sp. (D. Jollivet, pers. obs.) suggests the starting of a new<br />

site closed to an ancient vestimentiferan community. Those results illustrate the<br />

importance of regional dynamic on local processes.<br />

Latitudinal gradient and venting type explained less than half of the variation in the<br />

species data, suggesting that other local factors (i.e. biotic and abiotic factors) in<strong>de</strong>pen<strong>de</strong>nt<br />

of the type of venting may account for a large proportion of the variation observed. Small<br />

scale habitat heterogeneity (i.e. within a single animal aggregation) can highly influence<br />

species composition in response to species environmental tolerances and biotic<br />

interactions. Given the high variability in physical and chemical conditions at the<br />

centimetre scale within a single colony (Le Bris et al., 2006), even a small sampled area<br />

can integrate a wi<strong>de</strong> range of physico-chemical conditions. For example, at the Grommit<br />

site (21°33’S), two peltospirids species (i.e. Melanodrymia aurantiaca and Echinopelta<br />

fistulosa), usually associated with alvinellid colonies (Jollivet, 1996) were collected in<br />

290


Chapitre 5<br />

mussels from active black smokers’ chimney wall. On the other hand a sampling area<br />

smaller than a metre square may not be representative of the diversity of the whole<br />

community. Species distribution can vary between the centre and periphery of a single<br />

mussel bed (Bergquist et al., 2005) in response to high variability in environmental<br />

conditions within one metre (Sarrazin et al., 2006). In<strong>de</strong>ed, mobile gastropod species<br />

associated with alvinellid colonies, tubeworm clumps, bivalves and suspension-fee<strong>de</strong>rs, are<br />

not exclusive to a specific zone but rather occupy specific microhabitats that are present in<br />

more than one megafaunal community (Mills et al., 2007; Matabos et al., 2008a). Biotic<br />

interactions, including competition, predation and facilitation, may also play an important<br />

role on community structure.<br />

Limitations and issues<br />

The recent discovery of cryptic species following the emergence of bar-coding approaches<br />

raises the problem of species i<strong>de</strong>ntification and thus the assessment of the vent<br />

biodiversity. This appears to be a common issue in the sea. Ubiquist species are frequent<br />

occurrences in marine invertebrates and vertebrates and often found over a wi<strong>de</strong> range of<br />

latitu<strong>de</strong>s from the poles to the tropics (Knowlton, 1993). But good species i<strong>de</strong>ntification is<br />

nee<strong>de</strong>d to un<strong>de</strong>rstand the vent community history, colonisation pathways, and origin(s) of<br />

the fauna. Those observations reinforce the importance of comparative phylogeography to<br />

complement analytical approaches of historical biogeography (Arbogast & Kenagy, 2001).<br />

Correct assessments of species boundaries are fundamental to test biogeographic<br />

hypotheses (Palumbi, 1996). Furthermore, species i<strong>de</strong>ntifications are not always done by<br />

specialists, which lead to awkward comparisons between taxa lists. Valid comparisons<br />

across studies require a standardisation process for the i<strong>de</strong>ntification of organisms. This<br />

raises another issue related to species recording and species distribution areas and the<br />

validity of lists available in the literature. Individuals i<strong>de</strong>ntified by acknowledged<br />

specialists and recor<strong>de</strong>d in a museum with a vouch number appear to be a safe standard for<br />

biodiversity assessment and record. However, ecological studies are more numerous than<br />

systematics studies and lots of species lists are provi<strong>de</strong>d by ecologists (Van Dover, 2002,<br />

2003; Governar et al., 2005; Matabos et al. 2008a), so that it is still awkward to assess the<br />

real biodiversity associated with that part of the EPR. Differences in sampling methods<br />

(i.e. quantitative vs qualitative) can also account for the differences observed. Sampling<br />

<strong>de</strong>vices have been in<strong>de</strong>pen<strong>de</strong>ntly <strong>de</strong>veloped in the different labs and may thus not be<br />

comparable. Finally, the status of rare species also differs between systematic and<br />

291


Chapitre 5<br />

ecological studies and should be seriously consi<strong>de</strong>red in biogeographic approaches. Most<br />

ecological studies focus on a particular vent habitat, and may thus un<strong>de</strong>restimate the ‘real’<br />

local biodiversity. To our knowledge, the sediment habitat surrounding vents has been<br />

greatly neglected, so that many species may have never been collected.<br />

This study not only reinforces the existence of two biogeographic provinces along<br />

the East Pacific Rise separating the North from the South EPR, but also suggests that the<br />

‘abnormal’ high species diversity associated with the southern part of the SEPR (17°S-<br />

21°S) could result from an overlap of several biogeographic entities. Further comparisons<br />

are nee<strong>de</strong>d to confirm this assumption. Additional samples from the poorly explored<br />

Pacific Antarctic Ridge area could help in the un<strong>de</strong>rstanding of biodiversity patterns<br />

associated with the Pacific hydrothermal vent systems. Combining species diversity<br />

analyses with comparative phylogeography will help to unveil colonisation pathways that<br />

have led to the present species distribution. On the other hand, the lack of standards in<br />

species classification for <strong>de</strong>ep-sea biogeography and the cross-validation of species’<br />

i<strong>de</strong>ntification can explain the divergences observed when comparing studies conducted in<br />

the same parts of the ocean. The analysis of diversity associated with mussel beds was a<br />

first step towards a greater un<strong>de</strong>rstanding on the role of global, regional and local factors<br />

on biodiversity patterns. So far, combination of the latitudinal gradient and the venting<br />

significantly influenced the community structure but more replicates and explanatory<br />

variables are nee<strong>de</strong>d to ascertain such an assumption. This study will help <strong>de</strong>signing future<br />

macroecology studies along the EPR aiming to un<strong>de</strong>rstand processes influencing the<br />

distribution of biodiversity.<br />

AKNOWLEDGMENTS<br />

We thank the crew and pilots of the RV L’Atalante and the DSV Nautile for their<br />

assistance and technical support during the cruise BioSpeedo’04. We also thank the<br />

sequencing/genotyping platform GENOMER for mt COI sequences acquisition. This work<br />

was financially supported by the French programme Dorsales (INSU, Ifremer, CNRS), and<br />

the GDR Ecchis. A grant form the European SYNTHESIS Project fun<strong>de</strong>d the collaboration<br />

with the Swedish Museum of Natural History. This study also represents a contribution to<br />

the ANR Biodiversité ‘Deep Oases’ (ANR-06-BDIV-005).<br />

292


Chapitre 5<br />

REFERENCES<br />

Arbogast, B.S. & Kenagy, G.J. (2001) Comparative phylogeography as an integartive<br />

approach to historical biogeography. Journal of Biogeography, 28, 819-825.<br />

Auzen<strong>de</strong>, J.-M., Ballu, V., Batiza, R., Bi<strong>de</strong>au, D., Cormier, M.-H., Fouquet, Y.,<br />

Geistdoerfer, P., Lagabrielle, Y., Sinton, J. & Spa<strong>de</strong>a, P. (1994) Activité<br />

magmatique, tectonique et hydrothermale actuelle sur la dorsale est pacifique entre<br />

17°S et 19°S. (campagne naudur). Comptes Rendus <strong>de</strong> l'Académie <strong>de</strong>s Sciences,<br />

série II, 319, 811-818.<br />

Bachraty, C., Legendre, P. & Desbruyeres, D. (2009) Biogeographic relationships among<br />

<strong>de</strong>ep-sea hydrothermal vent faunas at global scale. Deep-Sea Research Part I<br />

Oceanographic Research Papers, 56, 1371-1378.<br />

Bergquist, D.C., Fleckenstein, C., Knisel, J., Begley, B., Macdonald, I.R. & Fisher, C.R.<br />

(2005) Variations in seep mussel bed communities along physical and chemical<br />

environmental gradients. Marine Ecology Progress Series, 293, 99-108.<br />

Briggs, J.C. (2006) Proximate sources of marine biodiversity. Journal of Biogeography,<br />

33, 1-10.<br />

Briggs, J.C. (2007) Marine biogeography and ecology: Invasions and introductions.<br />

Journal of Biogeography, 34, 193-198.<br />

Chevaldonné, P., Jollivet, D., Desbruyères, D., Lutz, R.A. & Vrijenhoek, R.C. (2002)<br />

Sister-species of eastern pacific hydrothermal vent worms (ampharetidae,<br />

alvinellidae, vestimentifera) provi<strong>de</strong> new mitochondrial coi clock calibration.<br />

Cahiers <strong>de</strong> Biologie Marine, 43, 367-370.<br />

Clarke, K.R. & Warwick, R.M. (2001) Change in marine communities: An approach to<br />

statistical analysis and interpretation, 2nd edition PRIMER-E Ltd edn, Plymouth,<br />

UK.<br />

De'ath, G. (2002) Mulitvariate regression trees: A new technique for mo<strong>de</strong>ling species -<br />

environment relationships. Ecology, 83, 1105-1117.<br />

Desbruyères, D., Hashimoto, J. & Fabri, M.-C. (2006a) Composition and biogeography of<br />

hydrothermal vent communities in western pacific back-arc basins. Geophysical<br />

monograph series, 166, 215-234.<br />

Desbruyères, D., Segonzac, M. & Bright, M. (2006b) Handbook of <strong>de</strong>ep-sea hydrothermal<br />

vent fauna., second edition edn. Denisia.<br />

Dreyer, J.C., Knick, K.E., Flickinger, W.B. & Van Dover, C.L. (2005) Development of<br />

macrofaunal community structure in mussel beds on the northern east pacific rise.<br />

Marine Ecology Progress Series, 302.<br />

Embley, R.W., Lupton, J.E., Massoth, G.J., Urabe, T., Tunnicliffe, V., Butterfield, D.A.,<br />

Shibata, T., Okano, O., Kinioshita, M. & Fujioka, K. (1998) Geological, chemical,<br />

and biological evi<strong>de</strong>nce for recent volcanism at 17.5°s: East pacific rise. Earth and<br />

Planetary Science Letters, 163, 131-147.<br />

Folmer, O., Black, M.B., Hoeh, W.R., Lutz, R.A. & Vrijenhoek, R.C. (1994) DNA primers<br />

for the amplification of mitochondrial cytochrome oxydase subunit 1 from diverse<br />

metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3, 294-299<br />

Fornari, D.J. & Embley, R.W. (1995) Tectonic and volcanic controls on hydrothermal<br />

processes at the mid-ocean ridge: An overview based on near-bottom and<br />

submersible studies. Seafloor hydrothermal systems (ed. by S.E. Humphris, R.A.<br />

Zierenberg, L.S. Mullineaux and R.E. Thomson). American Geophysical Union,<br />

Washington DC.<br />

293


Chapitre 5<br />

Fouquet, Y., Auzen<strong>de</strong>, J.-M., Ballu, V., Batiza, R., Bi<strong>de</strong>au, D., Cormier, M.-H.,<br />

Geistdoerfer, P., Lagabrielle, Y., Sinton, J. & Spa<strong>de</strong>a, P. (1994) Variabilité <strong>de</strong>s<br />

manifestations hydrothermales actuelles le long d'une dorsale ultra-rapi<strong>de</strong>: Exemple<br />

<strong>de</strong> la dorsale est pacifique entre 17°s et 19°s (campagne naudur). Comptes Rendus<br />

<strong>de</strong> l'Académie <strong>de</strong>s Sciences, série II, 319, 1399-1406.<br />

Fujioka, K., Massoth, G.J., De Ron<strong>de</strong>, C., Tanaka-Nakano, A., Gendron, J., Kinoshita, M.,<br />

Nishizawa, A., Okano, O., Paradis, G., Roe, K.J., Shinohara, M., Shitashima, K.,<br />

Tsunogai, U. & Ytow, N. (1995) Active hydrothermal systems of a super-fast<br />

spreading ridge, southern east pacific rise (13°58's to 18°26s). InterRidge News, 4,<br />

7-10.<br />

Geiger, D.L. & Thacker, C.E. (2005) Molecular phylogeny of vetigastropoda reveals nonmonophyletic<br />

scissurellidae, trochoi<strong>de</strong>a, and fissurelloi<strong>de</strong>a. Molluscan Research,<br />

25, 47-55.<br />

Geistdoerfer, P., Auzen<strong>de</strong>, J.-M., Batiza, R., Bi<strong>de</strong>au, D., Cormier, M.-H., Fouquet, Y.,<br />

Lagabrielle, Y., Sinton, J. & Spa<strong>de</strong>a, P. (1995) Hydrothermalisme et communautés<br />

animales associées sur la dorsale du pacifique oriental entre 17°s et 19°s (campagne<br />

naudur, décembre 1993). Comptes Rendus <strong>de</strong> l'Académie <strong>de</strong>s Sciences, série II,<br />

320, 47-54.<br />

Halanych, K.M., Tieger, M., O'mullan, G.D., Lutz, R.A. & Vrijenhoek, R.C. (1999) Brief<br />

<strong>de</strong>scription of biological communities at 7°s on the east pacific rise. International<br />

Ridge-Crest Research: Biological studies, 8, 23-27.<br />

Hawksworth, D.L. & Kalin-Arroyo, M.T. (1995) Magnitu<strong>de</strong> and distribution of<br />

biodiversity. Global biodiversity assessment (ed. by V.H. Heywood), pp. 107-192.<br />

University Press, Cambridge.<br />

Heads, M. (2005) Towards a panbiogeography of the seas. Biological Journal of the<br />

Linnean Society, 84, 675-723.<br />

Hurtado, L.A., Lutz, R.A. & Vrijenhoek, R.C. (2004) Distinct patterns of genetic<br />

differentiation among annelids of eastern pacific hydrothermal vents. Molecular<br />

Ecology, 13, 2603-2615.<br />

Hurtado, L.A., Mateos, M., Lutz, R.A. & Vrijenhoek, R.C. (2002) Molecular evi<strong>de</strong>nce for<br />

multiple species of oasisia (annelida: Siboglinidae) at eastern pacific hydrothermal<br />

vents. Cahier <strong>de</strong> Biologie Marine, 43, 377-380.<br />

Johnson, S.B., Warén, A. & Vrijenhoek, R.C. (2008) DNA barcoding of lepetodrilus<br />

limpets reveals cryptic species. Journal of Shellfish Research, 27, 43-51.<br />

Jollivet, D. (1996) Specific and genetic diversity at <strong>de</strong>ep-sea hydrothermal vents: An<br />

overview. Biodiversity and Conservation, 5, 1619-1653.<br />

Jollivet, D., Chevaldonné, P. & Planque, B. (1999) Hydrothermal vent alvinellid<br />

polychaete dispersal in the eastern pacific. 2. A metapopulation mo<strong>de</strong>l based on<br />

habitat shifts. Evolution, 53, 1128-1142.<br />

Jollivet, D. & Party, T.S.S. (2005) The biospeedo cruise: A new survey of hydrothermal<br />

vents along the south east pacific rise from 7°24s to 21°33s. InterRidge News, 13,<br />

20-26.<br />

Juniper, S.K. & Tunnicliffe, V. (1997) Crustal accretion and the hot vent ecosystem.<br />

Philosophical Transactions of the Royal Society of London. Series A, 355, 459-474.<br />

Knowlton, N. (1993) Sibling species in the sea. Ann. Rev. Ecol. Syst., 24, 189-216.<br />

Knowlton, N. (2000) Molecular genetic analyses of species boundaries in the sea.<br />

Hydrobiolgia, 420, 73-90.<br />

Le Bris, N., Govenar, B.W., Le Gall, C. & Fisher, C.R. (2006) Variability of physicochemical<br />

conditions in 9°50'n epr diffuse flow vent habitats. Marine Chemistry, 98,<br />

167-182.<br />

294


Chapitre 5<br />

Legendre, P. & Legendre, L. (1998) Numerical ecology. Elsevier.<br />

Levin, L.A., Etter, R.J., Rex, M.A., Gooday, J., Smith, C.R., Pineda, J., Stuart, C.T.,<br />

Hessler, R.R. & Pawson, D. (2001) Environmental influences on regional <strong>de</strong>ep-sea<br />

species diversity. Annual Review of Ecology and Systematics, 32, 51-93.<br />

Little, C.T.S. & Vrijenhoek, R.C. (2003) Are hydrothermal vent animals living fossils?<br />

Trends in ecology and evolution, 18, 582-588.<br />

Lonsdale, P. (1977) Clustering of suspension-feeding macrobenthos near abyssal<br />

hydrothermal vents at oceanic spreading centers. Deep Sea Research, 24, 857-863.<br />

Matabos, M., Le Bris, N., Pendlebury, S. & Thiébaut, E. (2008a) Role of physico-chemical<br />

environment on gastropod assemblages at hydrothermal vents on the east pacific<br />

rise (13°n/epr). Journal of the Marine Biological Association of the United<br />

Kingdom, 88, 995-1008.<br />

Matabos, M., Thiébaut, E., Le Guen, D., Sadosky, F., Jollivet, D. & Bonhomme, F.<br />

(2008b) Geographic clines and stepping-stone patterns <strong>de</strong>tected along the east<br />

pacific rise in the vetigastropod lepetodrilus elevatus reflect species crypticism.<br />

Marine Biology, 153, 545-563.<br />

Micheli, F., Peterson, C.H., Mullineaux, L.S., Fisher, C.R., Mills, S.W., Sancho, G.,<br />

Johnson, G.A. & Lenihan, H.S. (2002) Predation structures communties at <strong>de</strong>ep-sea<br />

hydrothermal vents. Ecological monographs, 72, 365-382.<br />

Mills, S.W., Mullineaux, L.S. & Tyler, P.A. (2007) Habitat associations in gastropod<br />

species at east pacific rise hydrothermal vents (9°50'n). Biological Bulletin, 212,<br />

185-194.<br />

Monge-Najera, J. (2008) Ecological biogeography: A review with emphasis on<br />

conservation and the neutral mo<strong>de</strong>l. Gayana, 72, 102-112.<br />

Mullineaux, L.S., Fisher, C.R., Peterson, C.H. & Schaeffer, S.W. (2000) Tubeworm<br />

succession at hydrothermal vents: Use of biogenic cues to reduce habitat selection<br />

error? Oecologia, 123, 275-284Mullineaux, L.S., Mills, S.W. & Goldman, E.<br />

(1998) Recruitment variation during a pilot colonization study of hydrothermal<br />

vents (9°50'n, east pacific rise). Deep Sea Research II, 45, 441-464.<br />

Mullineaux, L.S., Peterson, C.H., Micheli, F. & Mills, S.W. (2003) Successional<br />

mechanism varies along a gradient in hydrothermal fluid flux at <strong>de</strong>ep-sea vents.<br />

Ecological monographs, 73, 523-542.<br />

Palumbi, S.R. (1996) What can molecular genetics contribute to marine biogeography?<br />

Journal of Experimental Marine Biology and Ecology, 203, 75-92.<br />

Peek, A.S., Gustafson, R.G., Lutz, R.A. & Vrijenhoek, R.C. (1997) Evolutionary<br />

relationships of <strong>de</strong>ep-sea hydrothermal vent and cold-water seep clams (bivalvia:<br />

Vesicomyidae): Results from the mitochondrial cytochrome oxidase subunit i.<br />

Marine Biology, 130, 151-161.<br />

Sancho, G., Fisher, C.R., Mills, S.W., Micheli, F., Johnson, G.A., Lenihan, H.S., Peterson,<br />

C.H. & Mullineaux, L.S. (2005) Selective predation by the zoarcid fish thermarces<br />

cerberus at hydrothermal vents. Deep Sea Research I, 52, 837-844.<br />

Sarrazin, J. & Juniper, S.K. (1999) Biological characteristics of a hydrothermal edifice<br />

mosaic community. Marine Ecology Progress Series, 185, 1-19.<br />

Sarrazin, J., Juniper, S.K., Massoth, G.J. & Legendre, P. (1999) Physical and chemical<br />

factors influencing species distributions on hydrothermal sulfi<strong>de</strong> edifices of the<br />

juan <strong>de</strong> fuca ridge, northeast pacific. Marine Ecology Progress Series, 190, 89-112<br />

Sarrazin, J., Robigou, V., Juniper, S.K. & Delaney, J.R. (1997) Biological and geological<br />

dynamics over four years on a high-temperature sulfi<strong>de</strong> structure at the juan <strong>de</strong> fuca<br />

ridge hydrothermal observatory. Marine Ecology Progress Series, 153, 5-24.<br />

295


Chapitre 5<br />

Sarrazin, J., Walter, C., Sarradin, P.-M., Brind'amour, A., Desbruyères, D., Briand, P.,<br />

Fabri, M.-C., Van Gaever, S., Vanreusel, A., Bachraty, C. & Thiebaut, E. (2006)<br />

Community structure and temperature dynamics within a mussel assemblage in the<br />

southern east pacific rise. Cahiers <strong>de</strong> Biologie Marine, 47, 483-490.<br />

Shank, T.M., Fornari, D.J., Von Damn, K.L., Lilley, M.D., Haymon, R.M. & Lutz, R.A.<br />

(1998) Temporal and spatial patterns of biological community <strong>de</strong>velopment at<br />

nascent <strong>de</strong>ep-sea hydrothermal vents (9°50'n, east pacific rise). Deep Sea Research<br />

II, 45, 465-515.<br />

Spalding, M.D., Fox, H.E., Allen, G.R., Davidson, N., Ferdaña, Z.A., Finlayson, M.,<br />

Halpern, B.S., Jorge, M.A., Lombana, A., Lourie, S.A., Martin, K.D., Mcmanus,<br />

E., Omolnar, J., Recchia, C.A. & Robertson, J. (2007) Marine ecoregions of the<br />

world: A bioregionalization of coastal and shelf areas. BioScience, 57, 573-583.<br />

Stöhr, S. & Segonzac, M. (2006) Two new genera and species of ophiuroid<br />

(echino<strong>de</strong>rmata) from hydrothermal vents in the east pacific. Species diversity, 11,<br />

7-32.<br />

Tunnicliffe, V. (1988) Biogeography and evolution of hydrothermal-vent fauna in the<br />

eastern pacific ocean. Proceedings of the Royal Society of London B, 233, 347-366.<br />

Tunnicliffe, V. (1991) The biology of hydrothermal vents: Ecology and evolution.<br />

Oceanography and Marine Biology. An Annual Review, 29, 319-407.<br />

Tunnicliffe, V. & Fowler, C.M.R. (1996) Influence of sea-floor spreading on the global<br />

hydrothermal vent fauna. Nature, 379, 531-533.<br />

Tunnicliffe, V., Mcarthur, A.G. & Mchugh, D. (1998) A biogeographical perspective of<br />

the <strong>de</strong>ep-sea hydrothermal vent fauna. Advances in Marine Biology, 34, 353-442.<br />

Turnipseed, M., Knick, K.E., Lipcius, R.N., Dreyer, J.C. & Van Dover, C.L. (2003)<br />

Diversity in mussel beds at <strong>de</strong>ep-sea hydrothermal vents and cold seep. Ecology<br />

letters, 6, 518-523.<br />

Tyler, P.A., German, C.R., Ramirez-Llodra, E. & Van Dover, C.L. (2003) Un<strong>de</strong>rstanding<br />

the biogeography of chemosynthetic ecosystems. Oceanologica Acta, 25, 227-241.<br />

Van Dover, C., German, C.R., Speer, K.G., Parson, L.M. & Vrijenhoek, R.C. (2002)<br />

Evolution and biogeography of <strong>de</strong>ep-sea vent and seep invertebrates. Science, 295,<br />

1253-1257.<br />

Van Dover, C., Homphris, S.E., Fornari, D.J., Cavanaugh, C.M., Collier, R., Goffredi,<br />

S.K., Hashimoto, J., Lilley, M.D., Reysenbach, A.L., Shank, T., Von Damm, K.,<br />

Banta, A., Gallant, R.M. & Vrijenhoek, R.C. (2001a) Biogeography and ecological<br />

setting of indian ocean hydrothermal vents. Science, 294, 818-823.<br />

Van Dover, C. & Trask, J.L. (2000) Diversity at <strong>de</strong>ep-sea hydrothermal vent and intertidal<br />

mussel beds. Marine Ecology Progress Series, 195, 169-178.<br />

Van Dover, C.L. (2002) Community structure of mussels beds at <strong>de</strong>ep-sea hydrothermal<br />

vents. Marine Ecology Progress Series, 230, 137-158.<br />

Van Dover, C.L. (2003) Variation in community srtucture within hydrothermal vent<br />

mussel beds of the east pacific rise. Marine Ecology Progress Series, 253, 55-66.<br />

Van Dover, C.L. & Doerries, M.B. (2005) Community structure in mussel beds at<br />

logatchev hydrothermal vents and a comparison of macrofaunal species richness on<br />

slow- and fast-spreading mid-ocean ridges. Marine Ecology, 26, 110-120.<br />

Via, S. (2001) Sympatric speciation in animals: The ugly duckling grows up. Trends in<br />

Ecology & Evolution, 16, 381-390.<br />

Warén, A. & Bouchet, P. (2001) Gastropoda and monoplacophora from hydrothermal<br />

vents and seeps: New taxa and records. Veliger, 44, 116-231<br />

296


Chapitre 5<br />

Warwick, R.M. & Clarke, K.R. (2001) Practical measures of marine biodiversity based on<br />

relatedness of species. Oceanography and Marine Biology. An Annual Review, 39,<br />

207-231.<br />

Won, Y., Young, C.R., Lutz, R.A. & Vrijenhoek, R.C. (2003) Dispersal barriers and<br />

isolation among <strong>de</strong>ep-sea mussel populations (mytilidae: Bathymodiololus) from<br />

eastern pacific hydrothermal vents. Molecular Ecology, 12, 169-184.<br />

Young, C.R., Fujio, S. & Vrijenhoek, R.C. (2008) Directional dispersal between midocean<br />

ridges: Deep-ocean circulation and gene flow in ridgeia piscesae. Molecular<br />

Ecology, 17, 1718-1731.<br />

297


Chapitre 5<br />

298


Références Bibliographiques<br />

Références Bibliographiques<br />

A<br />

Akey JM, Eberle MA, Rie<strong>de</strong>r MJ, et al. (2004) Population history and natural selection<br />

shape patterns of genetic variation in 132 genes. Plos Biology 2, 1591-1599.<br />

Arbogast BS, Kenagy GJ (2001) Comparative phylogeography as an integartive approach<br />

to historical biogeography. Journal of Biogeography, 28, 819-825.<br />

Arellano SM, Young CM (2009) Spawning, <strong>de</strong>velopment and the duration of larval life in<br />

a <strong>de</strong>ep-sea cold-seep mussel. Biological Bulletin 216, 149-162.<br />

Auzen<strong>de</strong> J-M, Ballu V, Batiza, R, Bi<strong>de</strong>au D, Cormier M-H, Fouquet Y, Geistdoerfer P,<br />

Lagabrielle Y, Sinton J, Spa<strong>de</strong>a P (1994) Activité magmatique, tectonique et<br />

hydrothermale actuelle sur la dorsale est pacifique entre 17°S et 19°S. (campagne<br />

naudur). Comptes Rendus <strong>de</strong> l'Académie <strong>de</strong>s Sciences, série II, 319, 811-818.<br />

Asise JC (1987) I<strong>de</strong>ntification and interpretation of mitochondrial DNA stocks in marine<br />

species. In : Proceedings of stock the i<strong>de</strong>ntification Workshop. Kump L, Nakamura L<br />

(eds) National Oceanographic and Atmospheric Administation, Panama City, FL.<br />

Avise JC (1992) Molecular population structure and the biogeographic history of regional<br />

fauna: a case history with lessons for conservation biology. Oikos 63, 62-76.<br />

Avise JC (1994) Molecular markers, Natural History and Evolution. New York: Chapman<br />

& Hall.<br />

Avise JC (1998) The history and purview of phylogeography: a personal reflexion.<br />

Molecular Ecology 7, 371-379.<br />

Avise JC (2000) Phylogeography: the history and formation of species. Harvard<br />

University Press, Cambridge, MA, USA.<br />

Avise JC, Arnold J, Ball RM, et al. (1987) Intraspecific phylogeography: the<br />

mitochondrial DNA bridge between population genetics and systematics. Annual<br />

Review of Ecology and Systematics 18, 489-522.<br />

Avise JC, Ball RM (1990) Principles of genealogical concordance in species concepts and<br />

biological taxonomy. Oxford Survey of Evolutionnary Biology 7, 45-67.<br />

B<br />

Bachraty C, Legendre P, Desbruyeres D (2009) Biogeographic relationships among <strong>de</strong>epsea<br />

hydrothermal vent faunas at global scale. Deep-Sea Research Part I-<br />

Oceanographic Research Papers 56, 1371-1378.<br />

Baines JF, Chen Y, Das A, Stephan W (2002) DNA sequence variation at a duplicated<br />

gene: Excess of replacement polymorphism and extensive haplotype structure in<br />

the Drosophila melanogaster bicoid region. Molecular Biology and Evolution 19,<br />

989-998.<br />

Ban<strong>de</strong>lt HJ, Forster P, Rohl A (1999) Median-joining networks for inferring intraspecific<br />

phylogenies. Molecular Biology and Evolution 16, 37-48.<br />

Barton N, Bengtsson BO (1986) The Barrier to Genetic Exchange between Hybridizing<br />

Populations. Heredity 57, 357-376.<br />

Barton NH (1979) Gene flow past a cline. Heredity 43, 333-339.<br />

Barton NH (1986) The Effects of Linkage and Density-Depen<strong>de</strong>nt Regulation on Gene<br />

Flow. Heredity 57, 415-426.<br />

299


Références Bibliographiques<br />

Barton NH (2000) Estimating multilocus linkage disequilibria. Heredity 84, 373-389.<br />

Barton NH, Hewitt GM (1981) The Genetic-Basis of Hybrid Inviability in the Grasshopper<br />

Podisma-Pe<strong>de</strong>stris. Heredity 47, 367-383.<br />

Barton NH, Hewitt GM (1985) Analysis of Hybrid Zones. Annual Review of Ecology and<br />

Systematics 16, 113-148.<br />

Barton NH, Hewitt GM (1989) Adaptation, speciation and hybrid zones. Nature 341: 497-<br />

503.<br />

Bates AE, Tunnicliffe V, Lee RW (2005) Role of thermal conditions in habitat selection by<br />

hydrothermal vent gastropods. Marine Ecology-Progress Series 305, 1-15.<br />

Bazin E, Glemin S, Galtier N (2006) Population size does not influence mitochondrial<br />

genetic diversity in animals. Science 312, 570-572.<br />

Becker BJ, Fodrie FJ, McMillan PA, Levin LA (2005) Spatial and temporal variation in<br />

trace elemental fingerprints of mytilid mussel shells: A precursor to invertebrate<br />

larval tracking. Limnology and Oceanography 50, 48-61.<br />

Becker BJ, Levin LA, Fodrie FJ, McMillan PA (2007) Complex larval connectivity<br />

patterns among marine invertebrate populations. Proceedings of the National<br />

Aca<strong>de</strong>my of Sciences of the United States of America 104, 3267-3272.<br />

Beerli P, Felsenstein J (1999) Maximum-likelihood estimation of migration rates and<br />

effective population numbers in two populations using a coalescent approach.<br />

Genetics 152: 763-773.<br />

Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and<br />

effective population sizes in n subpopulations by using a coalescent approach.<br />

Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of America 98,<br />

4563-4568.<br />

Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and<br />

effective population sizes in n subpopulations by using a coalescent approach.<br />

Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of America<br />

98, 4563-4568.<br />

Belkhir K, Borsa P, Chikhi L, Raufaste N, Bonhomme F (2004) Genetix 4.05, logiciel sous<br />

Windows TM <strong>pour</strong> la génétique <strong>de</strong>s populations. Laboratoire Génome, Populations,<br />

Interactions, Adaptations, UMR 5000, Université <strong>de</strong> Montpellier 2. Available at:<br />

www.genetix.univ-montp2.fr/ genetix/genetix.htm<br />

Bengtsson J (1989) Interspecific Competition Increases Local Extinction Rate in a<br />

Metapopulation System. Nature 340, 713-715.<br />

Bergquist DC, Fleckenstein C, Knisel J, Begley B, Macdonald IR, Fisher CR (2005)<br />

Variations in seep mussel bed communities along physical and chemical<br />

environmental gradients. Marine Ecology Progress Series, 293, 99-108.<br />

Bermingham E, Avise JC (1986) Molecular Zoogeography of Fresh-Water Fishes in the<br />

Southeastern United-States. Genetics 113, 939-965.<br />

Bermingham E, Martin AP (1998) Comparative mtDNA phylogeography of neotropical<br />

freshwater fishes: testing shared history to infer the evolutionary landscape of<br />

lower Central America. Molecular Ecology 7, 499-517.<br />

Berry AJ, Ajioka JW, Kreitman M (1991) Lack of Polymorphism on the Drosophila 4th<br />

Chromosome Resulting from Selection. Genetics 129, 1111-1117.<br />

Bert TM, Arnold WS (1995) An Empirical-Test of Predictions of 2 Competing Mo<strong>de</strong>ls for<br />

the Maintenance and Fate of Hybrid Zones - Both Mo<strong>de</strong>ls Are Supported in a<br />

Hard-Clam Hybrid Zone. Evolution 49, 276-289.<br />

Bert TM, Harrison RG (1988) Hybridization in Western Atlantic Stone Crabs (Genus<br />

Menippe) - Evolutionary History and Ecological Context Influence Species<br />

Interactions. Evolution 42, 528-544.<br />

300


Références Bibliographiques<br />

Bierne N, Borsa P, Daguin C, et al. (2003) Introgression patterns in the mosaic hybrid zone<br />

between Mytilus edulis and M-galloprovincialis. Molecular Ecology 12, 447-461.<br />

Bierne N, Tanguy A, Faure M, et al. (2007) Mark-recapture cloning: a straightforward and<br />

cost-effective cloning method for population genetics of single copy nuclear DNA<br />

sequences in diploids. Molecular Ecology Notes.<br />

Birky CW (1995) Uniparental Inheritance of Mitochondrial and Chloroplast Genes -<br />

Mechanisms and Evolution. Proceedings of the National Aca<strong>de</strong>my of Sciences of the<br />

United States of America 92, 11331-11338.<br />

Birky CW (2001) The inheritance of genes in mitochondria and chloroplasts: Laws,<br />

mechanisms, and mo<strong>de</strong>ls. Annual Review of Genetics 35, 125-148.<br />

Birky CW, Fuerst P, Maruyama T (1989) Organelle Gene Diversity un<strong>de</strong>r Migration,<br />

Mutation, and Drift - Equilibrium Expectations, Approach to Equilibrium, Effects of<br />

Heteroplasmic Cells, and Comparison to Nuclear Genes. Genetics 121, 613-627.<br />

Birky CW, Maruyama T, Fuerst P (1983) An Approach to Population and Evolutionary<br />

Genetic Theory for Genes in Mitochondria and Chloroplasts, and Some Results.<br />

Genetics 103, 513-527.<br />

Black MB, Lutz RA, Vrijenhoek RC (1994) Gene Flow among Vestimentiferan Tube<br />

Worm (Riftia pachyptila) Populations from Hydrothermal Vents of the Eastern<br />

Pacific. Marine Biology 120, 33-39.<br />

Bodmer WF, Cavalli-Sforza LL (1968) A migration matrix mo<strong>de</strong>l for the study of genetic<br />

drift. Genetics 59, 565-592.<br />

Bohrmann G, Chin C, Petersen S, et al. (1999) Hydrothermal activity at Hook Ridge in the<br />

Central Bransfield Basin, Antarctica. Geo-Marine Letters 18, 277-284.<br />

Boon E, Faure MF, Bierne N (2009) The Flow of Antimicrobial Pepti<strong>de</strong> Genes Through a<br />

Genetic Barrier Between Mytilus edulis and M-galloprovincialis. Journal of<br />

Molecular Evolution 68, 461-474.<br />

Boutet I, Tanguy A, Le Guen D, Piccino P, Hour<strong>de</strong>z S, Legendre P, Jollivet D (2009)<br />

Global <strong>de</strong>pression in gene expression as a response to rapid thermal changes in vent<br />

mussels. Proceedings of the Royal Society of London Series B.<br />

doi:10.1098/rspb.2009.0503<br />

Breton S, Beaupre HD, Stewart DT, Hoeh WR, Blier PU (2007) The unusual system of<br />

doubly uniparental inheritance of mtDNA: isn't one enough? Trends in Genetics 23,<br />

465-474.<br />

Briggs JC (2006) Proximate sources of marine biodiversity. Journal of Biogeography, 33,<br />

1-10.<br />

Briggs JC (2007) Marine biogeography and ecology: Invasions and introductions. Journal<br />

of Biogeography, 34, 193-198.<br />

Broughton RE, Harrison RG (2003) Nuclear gene genealogies reveal historical,<br />

<strong>de</strong>mographic and selective factors associated with speciation in field crickets. Genetics<br />

163, 1389-1401.<br />

Burton RS (1998) Intraspecific phylogeography across the Point Conception biogeographic<br />

boundary. Evolution 52, 734-745.<br />

Burton RS, Lee BN (1994) Nuclear and Mitochondrial Gene Genealogies and Allozyme<br />

Polymorphism across a Major Phylogeographic Break in the Copepod Tigriopus-<br />

Californicus. Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States of<br />

America 91, 5197-5201.<br />

301


Références Bibliographiques<br />

C<br />

Cannon GA, Pashinski DJ (1997) Variation in mean currents affecting hydrothermal<br />

plumes of the Juan <strong>de</strong> Fuca Ridge. Journal of Geophysical Research 102: 24965-<br />

24976.<br />

Cao LQ, Kenchington E, Zouros E (2004) Differential segregation patterns of sperm<br />

mitochondria in embryos of the blue mussel (Mytilus edulis). Genetics 166, 883-894.<br />

Carling MD, Brumfield RT (2007) Gene sampling strategies for multi-locus population<br />

estimates of genetic diversity (theta). PLoS One 2, e160.<br />

Carney SL, Formica MI, Divatia H, Nelson K, Fisher CR, Schaeffer SW (2006) Population<br />

structure of the mussel “Bathymodiolus” childressi from Gulf of Mexico hydrocarbon<br />

seeps. Deep-Sea Research I 53: 1061-1072.<br />

Castric V, Bechsgaard J, Schierup MH, Vekemans X (2008) Repeated Adaptive<br />

Introgression at a Gene un<strong>de</strong>r Multiallelic Balancing Selection. Plos Genetics 4.<br />

Castric V, Vekemans X (2007) Evolution un<strong>de</strong>r strong balancing selection: how many<br />

codons <strong>de</strong>termine specificity at the female self-incompatibility gene SRK in<br />

Brassicaceae? Bmc Evolutionary Biology 7.<br />

Charlesworth B, Morgan MT, Charlesworth D (1993) The effect of <strong>de</strong>leterious mutations<br />

on neutral molecular variation. Genetics 134, 1289-1303.<br />

Chevaldonné P, Desbruyeres D, Childress JJ (1992) Some Like Hot - and Some Even<br />

Hotter. Nature 359, 593-594.<br />

Chevaldonné P, Jollivet D (1993) Vi<strong>de</strong>oscopic study of <strong>de</strong>ep-sea hydrothermal vent<br />

alvinellid polychaete populations: biomass estimation and behaviour. Marine Ecology<br />

Progress Series 95, 251-262.<br />

Chevaldonné P, Jollivet D, Desbruyeres D, Lutz RA, Vrijenhoek RC (2002) Sister-species<br />

of eastern Pacific hydrothermal vent worms (Ampharetidae, Alvinellidae,<br />

Vestimentifera) provi<strong>de</strong> new mitochondrial COI clock calibration. Cahiers De<br />

Biologie Marine 43, 367-370.<br />

Chevaldonné P, Jollivet D, Vangriesheim A, Desbruyeres D (1997) Hydrothermal-vent<br />

alvinellid polychaete dispersal in the eastern Pacific .1. Influence of vent site<br />

distribution, bottom currents, and biological patterns. Limnology and Oceanography<br />

42, 67-80.<br />

Childress JJ, Fisher CR (1992) The biology of hydrothermal vent animals: physiology,<br />

biochemistry and autotrophic symbioses. Oceanography and Marine Biology. An<br />

annual review 30, 337-441.<br />

Clarke KR, Warwick RM (2001) Change in marine communities: An approach to<br />

statistical analysis and interpretation, 2nd edition PRIMER-E Ltd edn, Plymouth,<br />

UK.<br />

Collin R (2001) The effects of mo<strong>de</strong> of <strong>de</strong>velopment on phylogeography and population<br />

structure of North Atlantic Crepidula (Gastropoda: Calyptraeidae). Molecular<br />

Ecology 10, 2249-2262.<br />

Comtet T, Desbruyeres D (1998) Population structure and recruitment in mytilid bivalves<br />

from the Lucky Strike and Menez Gwen hydrothermal vent fields (37 <strong>de</strong>grees 17 ' N<br />

and 37 <strong>de</strong>grees 50 ' N on the Mid-Atlantic Ridge). Marine Ecology-Progress Series<br />

163, 165-177.<br />

Comtet T, Le Pennec M, Desbruyeres D (1999) Evi<strong>de</strong>nce of a sexual pause in<br />

Bathymodiolus azoricus (Bivalvia : Mytilidae) from hydrothermal vents of the Mid-<br />

Atlantic Ridge. Journal of the Marine Biological Association of the United Kingdom<br />

79, 1149-1150.<br />

302


Références Bibliographiques<br />

Cormier M-H (1997) The ultrafast East Pacific Rise: instability of the plate boundary and<br />

implications for accretionary processes. Philosophical Transactions of the Royal<br />

Society of London 355, 341-367.<br />

Cormier M-H (1997) The ultrafast East Pacific Rise: instability of the plate boundary and<br />

implications for accretionary processes. Philosophical Transactions of the Royal<br />

Society of London 355, 341-367.<br />

Cowen JP, Fornari DJ, Shank TM, et al. (2007) Volcanic Eruptions at East Pacific Rise<br />

near 9°50’N. EOS, Transactions of the American Geophysical Union 88, 81-83.<br />

Cowen JP, Fornari DJ, Shank TM, et al. (2007) Volcanic Eruptions at East Pacific Rise<br />

near 9°50’N. EOS, Transactions of the American Geophysical Union 88, 81-83.<br />

Cox BC (2001) The biogeographic regions reconsi<strong>de</strong>red. Journal of Biogeography 28,<br />

511-523.<br />

Cox CB, Moore PD (2000) Drawing lines in the water. In Biogeography : an ecological<br />

and evolutionary approach. Blackwell Science, London, 238-261.<br />

Craddock C, Hoeh WR, Lutz RA, Vrijenhoek RC (1995) Extensive Gene Flow among<br />

Mytilid (Bathymodiolus thermophilus) Populations from Hydrothermal Vents of the<br />

Eastern Pacific. Marine Biology 124, 137-146.<br />

Craddock C, Hoeh WR, Lutz RA, Vrijenhoek RC (1995) Extensive Gene Flow among<br />

Mytilid (Bathymodiolus-Thermophilus) Populations from Hydrothermal Vents of the<br />

Eastern Pacific. Marine Biology 124, 137-146.<br />

Craddock C, Lutz RA, Vrijenhoek RC (1997) Patterns of dispersal and larval <strong>de</strong>velopment<br />

of archaeogastropod limpets at hydrothermal vents in the eastern Pacific. Journal of<br />

Experimental Marine Biology and Ecology 210, 37-51.<br />

Cunha RL, Castilho R, Ruber L, Zardoya R (2005) Patterns of cladogenesis in the<br />

venomous marine gastropod genus Conus from the Cape Ver<strong>de</strong> islands. Systematic<br />

Biology 54, 634-650.<br />

Cunningham CW, Collins TM (1994) Developing mo<strong>de</strong>l systems for molecular<br />

biogeography : vicariance and interchange in marine invertebrates. Wagner GP,<br />

DeSalle R (eds). In : Molecular ecology and evolution : approaches and applications.<br />

Birkhauser Velag, Basel, Switzerland.<br />

Cunningham CW, Collins TM (1998) Beyond area relationships : extinction and<br />

recolonization in molecular marine biogeography. In : Molecular Approaches to<br />

Ecology and Evolution. DeSalle R, Schierwater B (eds), Birkhäser Verlag Basel, 297-<br />

321.<br />

D<br />

Dalziel AC, Stewart DT (2002) Tissue-specific expression of male-transmitted<br />

mitochondrial DNA and its implications for rates of molecular evolution in Mytilus<br />

mussels (Bivalvia : Mytilidae). Genome 45, 348-355.<br />

Dawson MN (2001) Phylogeography in coastal marine animals: a solution from<br />

California? Journal of Biogeography 28, 723-736.<br />

De'ath G (2002) Mulitvariate regression trees: A new technique for mo<strong>de</strong>ling species -<br />

environment relationships. Ecology, 83, 1105-1117.<br />

Desbruyères D, Alayse-Danet AM, Ohta S (1994) Deep-sea hydrothermal communities in<br />

southwestern Pacific back-arc basins (the North Fiji and Lau Basins): Composition,<br />

microdistribution and food web. Marine Geology 116, 227-242.<br />

303


Références Bibliographiques<br />

Desbruyères D, Almeida A, Biscoito M, et al. (2000) Distribution of hydrothermal vent<br />

communities along the Northern Mid-Atlantic Ridge. Dispersal vs. Environmental<br />

control : a review. Hydrobiologia 440, 201-216.<br />

Desbruyeres D, Chevaldonné P, Alayse AM, et al. (1998) Biology and ecology of the<br />

"Pompeii worm" (Alvinella pompejana Desbruyeres and Laubier), a normal dweller of<br />

an extreme <strong>de</strong>ep-sea environment: A synthesis of current knowledge and recent<br />

<strong>de</strong>velopments. Deep-Sea Research Part Ii-Topical Studies in Oceanography 45, 383<br />

Desbruyeres D, Hashimoto J, Fabri MC (2006a) Composition and biogeography of<br />

hydrothermal vent communities in western pacific back-arc basins. Back-Arc<br />

Spreading Systems: Geological, Biological, Chemical, and Physical Interactions 166,<br />

215-234.<br />

Desbruyeres D, Laubier L (1991) Systematics, Phylogeny, Ecology and Distribution of the<br />

Alvinellidae (Polychaeta) from Deep-Sea Hydrothermal Vents. Ophelia, 31-45.<br />

Desbruyères D, Segonzac M, Bright M (2006b) Handbook of <strong>de</strong>ep-sea hydrothermal vent<br />

fauna. Desbruyères D, Segonzac M, Bright M (eds), Denisia 18, 516-517.<br />

Dixon DR, Lowe DM, Miller PI, et al. (2006) Evi<strong>de</strong>nce of seasonal reproduction in the<br />

Atlantic vent mussel Bathymodiolus azoricus, and an apparent link with the timing<br />

of photosynthetic primary production. Journal of the Marine Biological<br />

Association of the United Kingdom 86, 1363-1371.<br />

Donoghue MJ, Moore BR (2003) Toward an integrative historical biogeography.<br />

Integrative and Comparative Biology 43, 261-270.<br />

Dooh RT, Adamowicz SJ, Hebert PDN (2006) Comparative phylogeography of two North<br />

American 'glacial relict' crustaceans. Molecular Ecology 15, 4459-4475.<br />

Doyle JJ, Dickson E (1987) Preservation of plant samples for DNA restriction<br />

endonuclease analysis. Taxon 36, 715–722.<br />

Doyle RF (1985) Biogeographical studies of rocky shores near Point Conception,<br />

California. PhD dissertation, University of California, Santa Barbara, CA.<br />

Dreyer JC, Knick KE, Flickinger WB, Van Dover CL (2005) Development of macrofaunal<br />

community structure in mussel beds on the northern east pacific rise. Marine<br />

Ecology Progress Series, 302.<br />

Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling<br />

trees. BMC Evolutionary Biology 7 doi:10.1186/1471-2148-7-214.<br />

Duda TF, Palumbi SR (1999) Population structure of the black tiger prawn, Penaeus<br />

monodon, among western Indian Ocean and western Pacific populations. Marine<br />

Biology 134, 705-710.<br />

E<br />

Eakins B, Lonsdale P (2003) Structural patterns and tectonic history of the Bauer<br />

microplate, Eastern Tropical Pacific. Marine Geophysical Researches 24, 171-205.<br />

Embley RW, Lupton JE, Massoth GJ, Urabe T, Tunnicliffe V, Butterfield DA, Shibata T,<br />

Okano O, Kinioshita M, Fujioka K (1998) Geological, chemical, and biological<br />

evi<strong>de</strong>nce for recent volcanism at 17.5°s: East pacific rise. Earth and Planetary<br />

Science Letters, 163, 131-147.<br />

Edwards SV, Beerli P (2000) Perspective: Gene divergence, population divergence, and<br />

the variance in coalescence time in phylogeographic studies. Evolution 54, 1839-1854.<br />

Emerson BC, Paradis E, Thebaud C (2001) Revealing the <strong>de</strong>mographic histories of species<br />

using DNA sequences. Trends in Ecology & Evolution 16, 707-716.<br />

304


Références Bibliographiques<br />

Emsley MG (1965) Speciation in Helioconus (Lep. Nymphalidae): morphology and<br />

geographic distribution. Zoologica (NY) 50, 191-254.<br />

Endler J (1982) Pleistocen forest refuges : fact or fancy. In Biological diversification in the<br />

Tropics. Prance GT (ed), Colombia University Press, New York, 641-657.<br />

Engle VD, Summers JK (1999) Latitudinal gradients in benthic community composition in<br />

Western Atlantic estuaries. Journal of Biogeography 26, 1007-1023.<br />

Excoffier L, Laval G, Schnei<strong>de</strong>r S (2005) Arlequin (version 3.0): An integrated software<br />

package for population genetics data analysis. Evolutionary Bioinformatics 1, 47-50.<br />

F<br />

Faure B (2009) Etu<strong>de</strong> multi-locus <strong>de</strong>s pressions évolutives s’excerçant sur <strong>de</strong>s bivalves<br />

marins en situation d’hybridation : cas <strong>de</strong>s moules hydrothermales du genre<br />

Bathymodiolus <strong>de</strong>s dorsales médio-atlantiques et est-pacifique. Thèse <strong>de</strong> Doctorat,<br />

Université Montpellier 2.<br />

Faure B, Chevaldonné P, Pradillon F, Thiébaut E, Jollivet D (2007) Spatial and temporal<br />

dynamics of reproduction and settlement in the Pompeii worm Alvinella pompejana<br />

(Polychaeta: Alvinellidae). Marine Ecology Progress Series 348, 197-211.<br />

Faure B, Jollivet D, Tanguy A, Bonhomme F, Bierne N (2009) Speciation in the <strong>de</strong>ep sea:<br />

multi-locus analysis of divergence and gene flow between two hybridizing species of<br />

hydrothermal vent mussels. PLoS One 4, e6485.<br />

Faure MF, David P, Bonhomme F, Bierne N (2008) Genetic hitchhiking in a subdivi<strong>de</strong>d<br />

population of Mytilus edulis. Bmc Evolutionary Biology 8.<br />

Felsenstein J (1981) Evolutionary trees from DNA sequences: a maximum likelihood<br />

approach. Journal of Molecular Evolution 17, 368-376.<br />

Felsenstein J (2006) Accuracy of coalescent likelihood estimates: Do we need more sites,<br />

more sequences, or more loci? Molecular Biology and Evolution 23, 691-700.<br />

Feral JP (2002) How useful are the genetic markers in attempts to un<strong>de</strong>rstand and manage<br />

marine biodiversity? Journal of Experimental Marine Biology and Ecology 268,<br />

121-145.<br />

Fialamedioni A, Metivier C, Herry A, Lepennec M (1986) Ultrastructure of the Gill of the<br />

Hydrothermal-Vent Mytilid Bathymodiolus Sp. Marine Biology 92, 65-72.<br />

Filatov DA, Charlesworth D (1999) DNA polymorphism, haplotype structure and<br />

balancing selection in the leavenworthia PgiC locus. Genetics 153, 1423-1434.<br />

Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification<br />

of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates.<br />

Molecular Marine Biology and Biotechnology 3, 294-299.<br />

Fornari DJ, Embley RW (1995) Tectonic and volcanic controls on hydrothermal processes<br />

at the mid-ocean ridge: An overview based on near-bottom and submersible<br />

studies. Seafloor hydrothermal systems (ed. by S.E. Humphris, R.A. Zierenberg,<br />

L.S. Mullineaux and R.E. Thomson). American Geophysical Union, Washington<br />

DC.<br />

Fouquet Y, Auzen<strong>de</strong> JM, Ballu V, et al. (1994) Variability of the Present-Day<br />

Hydrothermal Activity Along an Ultra-Fast Spreading Ridge - East Pacific Rise<br />

between 17-Degrees-C and 19-Degrees-S (Naudur Cruise). Comptes Rendus De L<br />

Aca<strong>de</strong>mie Des Sciences Serie Ii 319, 1399-1406.<br />

France SC, Hessler RR, Vrijenhoek RC (1992) Genetic differentiation between spatiallydisjunct<br />

populations of the <strong>de</strong>ep-sea, hydrothermal vent-en<strong>de</strong>mic amphipod Ventiella<br />

sulfuris. Marine Biology 114, 551-559.<br />

305


Références Bibliographiques<br />

Francheteau J, Armijo R, Cheminee JL, et al (1990) 1 Ma East Pacific Rise Oceanic-Crust<br />

and Uppermost Mantle Exposed by Rifting in Hess Deep (Equatorial Pacific-Ocean).<br />

Earth and Planetary Science Letters 101: 281-295.<br />

Francheteau J, Armijo R, Cheminee JL, et al. (1990) 1 Ma East Pacific Rise Oceanic-Crust<br />

and Uppermost Mantle Exposed by Rifting in Hess Deep (Equatorial Pacific-Ocean).<br />

Earth and Planetary Science Letters 101, 281-295.<br />

Fretter V (1988) New archaeogastropod limpets from hydrothermal vents; superfamily<br />

Lepetodrilacea. II. Anatomy. Philosophical Transactions of the Royal Society of<br />

London Series B 318, 33-82.<br />

Fu YX, Li WH (1993) Statistical test of neutrality of mutations. Genetics 133: 693-709.<br />

Fujio SZ, Imasato N (1991) Diagnostic Calculation for Circulation and Water Mass<br />

Movement in the Deep Pacific. Journal of Geophysical Research-Oceans 96, 759-774.<br />

Fujioka K, Massoth GJ, De Ron<strong>de</strong> C, Tanaka-Nakano A, Gendron J, Kinoshita M,<br />

Nishizawa A, Okano O, Paradis G, Roe KJ, Shinohara M, Shitashima K, Tsunogai<br />

U, Ytow N (1995) Active hydrothermal systems of a super-fast spreading ridge,<br />

southern east pacific rise (13°58's to 18°26s). InterRidge News, 4, 7-10.<br />

Fukushima T (2000) Deep sea benthology : history and present state. MBRIJ Annual<br />

Report, 74-82.<br />

Fustec A, Desbruyères D, Juniper SK (1987) Genetic differentiation between spatiallydisjunct<br />

populations of the <strong>de</strong>ep-sea, hydrothermal vent-en<strong>de</strong>mic amphipod Ventiella<br />

sulfuris. Biological Oceanography 4, 121-164.<br />

G<br />

Gagnière N, Jollivet D, Boutet I, et al (submitted) Insights metazoan evolution from<br />

Alvinella pompejana cDNAs. Genome Research.<br />

Galtier N, Depaulis F, Barton NH (2000) Detecting bottlenecks and selective sweeps from<br />

DNA sequence polymorphism. Genetics 155, 981-987.<br />

Garnier-Géré P, Dillmann C (1992) A Computer-Program for Testing Pairwise Linkage<br />

Disequilibria in Subdivi<strong>de</strong>d Populations. Journal of Heredity 83, 239-239.<br />

Geiger DL, Thacker CE (2005) Molecular phylogeny of vetigastropoda reveals nonmonophyletic<br />

scissurellidae, trochoi<strong>de</strong>a, and fissurelloi<strong>de</strong>a. Molluscan Research,<br />

25, 47-55.<br />

Geistdoerfer P, Auzen<strong>de</strong> J-M, Batiza R, Bi<strong>de</strong>au D, Cormier M-H, Fouquet Y, Lagabrielle<br />

Y, Sinton J, Spa<strong>de</strong>a P (1995) Hydrothermalisme et communautés animales<br />

associées sur la dorsale du pacifique oriental entre 17°s et 19°s (campagne naudur,<br />

décembre 1993). Comptes Rendus <strong>de</strong> l'Académie <strong>de</strong>s Sciences, série II, 320, 47-54.<br />

Glémin S, Gau<strong>de</strong> T, Guillemin ML, et al. (2005) Balancing selection in the wild: Testing<br />

population genetics theory of self-incompatibility in the rare species Brassica insularis.<br />

Genetics 171, 279-289.<br />

Glinka S, Ometto L, Mousset S, Stephan W, De Lorenzo D (2003) Demography and<br />

natural selection have shaped genetic variation in Drosophila melanogaster: A<br />

multi-locus approach. Genetics 165, 1269-1278.<br />

Godinho R, Crespo EG, Ferrand N (2008) The limits of mtDNA phylogeography: complex<br />

patterns of population history in a highly structured Iberian lizard are only revealed by<br />

the use of nuclear markers. Molecular Ecology 17, 4670-4683.<br />

Goldstien SJ, Schiel DR, Gemmell NJ (2006) Comparative phylogeography of coastal<br />

limpets across a marine disjunction in New Zealand. Molecular Ecology 15, 3259-<br />

3268.<br />

306


Références Bibliographiques<br />

Graham JH, Freeman DC, McArthur ED (1995) Narrow Hybrid Zone between 2<br />

Subspecies of Big Sagebrush (Artemisia-Tri<strong>de</strong>ntata, Asteraceae) .2. Selection<br />

Gradients and Hybrid Fitness. American Journal of Botany 82, 709-716.<br />

Grassle JP (1985) Genetic differentiation in populations of hydrothermal vent mussels<br />

(Bathymodiolus thermophilus) from the Galapagos Rift and 13°N on the East Pacific<br />

Rise. Biological Society of Washington 6, 429-442.<br />

Graur D, Li WH (1991) Neutral Mutation Hypothesis Test. Nature 354, 114-115.<br />

Griffiths RC, Tavare S (1994a) Sampling Theory for Neutral Alleles in a Varying<br />

Environment. Philosophical Transactions of the Royal Society of London Series B-<br />

Biological Sciences 344, 403-410.<br />

Griffiths RC, Tavare S (1994b) Simulating Probability-Distributions in the Coalescent.<br />

Theoretical Population Biology 46, 131-159.<br />

H<br />

Haddrill PR, Thornton KR, Charlesworth B, Andolfatto P (2005) Multilocus patterns of<br />

nucleoti<strong>de</strong> variability and the <strong>de</strong>mographic and selection history of Drosophila<br />

melanogaster populations. Genome Research 15, 790-799.<br />

Haffer J (1969) Speciation in Amozonian forest birds. Science 165, 131-136.<br />

Halanych KM, Tieger M, O'mullan GD, Lutz RA, Vrijenhoek RC (1999) Brief <strong>de</strong>scription<br />

of biological communities at 7°s on the east pacific rise. International Ridge-Crest<br />

Research: Biological studies 8, 23-27.<br />

Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis<br />

program for Windows 95/98/NT. Nucleic Acids Symposium Series 41, 95-98.<br />

Hannington MD, Jonasson IR (1995) Physical and chemical processes of seafloor<br />

mineralization at mid-ocean ridges. In Seafloor Hydrothermal Systems. Humphris,<br />

SE, Zierenberg, RA, Mullineaux, LS, Thompson, RE (eds). Geophysical<br />

Monograph. American Geophysical Society Washington, 91: 115-157.<br />

Harpending HC, Batzer MA, Gurven M, et al. (1998) Genetic traces of ancient<br />

<strong>de</strong>mography. Proceedings of the National Aca<strong>de</strong>my of Sciences of the United States<br />

of America 95, 1961-1967.<br />

Harr B, Kauer M, Schlotterer C (2002) Hitchhiking mapping: A population-based finemapping<br />

strategy for adaptive mutations in Drosophila melanogaster. Proceedings<br />

of the National Aca<strong>de</strong>my of Sciences of the United States of America 99, 12949-<br />

12954.<br />

Harrison RG (1990) Hybrid zones: windows on evolutionary processes. Oxf. Surv. Evol.<br />

Biol. 7: 69-128.<br />

Harrison S, Hastings A (1996) Genetic and evolutionary consequences of metapopulation<br />

structure. Trends in Ecology & Evolution 11, 180-183.<br />

Hartl DL, Clark AG (2007) Inbreeding, Population, and Migration. In Principes of<br />

population genetics, Sinauer AD (ed), Sinauer, 257-315.<br />

Hasegawa M, Kishino H, Yano T (1985) Dating the human-ape splitting by a molecular<br />

clock of mitochondrial DNA. Journal of Molecular Evolution 22: 160-174.<br />

Hashimoto J, Ohta S, Gamo T, et al (2001) Hydrothermal vents and associated biological<br />

communities in the Indian Ocean. InterRidge News 10 : 21-22.<br />

Hawksworth DL, Kalin-Arroyo MT (1995) Magnitu<strong>de</strong> and distribution of biodiversity.<br />

Global biodiversity assessment (ed. by V.H. Heywood), pp. 107-192. University<br />

Press, Cambridge.<br />

307


Références Bibliographiques<br />

Haymon RM, Fornari DJ, Edwards MH, et al. (1991) Hydrothermal Vent Distribution<br />

Along the East Pacific Rise Crest (9-Degrees-09'-54'n) and Its Relationship to<br />

Magmatic and Tectonic Processes on Fast-Spreading Midocean Ridges. Earth and<br />

Planetary Science Letters 104, 513-534.<br />

Haymon RM, Fornari DJ, Vondamm KL, et al. (1993) Volcanic-Eruption of the Midocean<br />

Ridge Along the East Pacific Rise Crest at 9-Degrees-45-52'n - Direct<br />

Submersible Observations of Sea-Floor Phenomena Associated with an Eruption<br />

Event in April, 1991. Earth and Planetary Science Letters 119, 85-101.<br />

Heads M (2005) Towards a panbiogeography of the seas. Biological Journal of the<br />

Linnean Society, 84, 675-723.<br />

Hessler RR, Lonsdale PF (1991) Biogeography of Mariana Trough hydrothermal vents.<br />

Deep Sea Research 38, 185-199.<br />

Hessler RR, Smithey Jr WM, Keller CH (1985) Spatial and temporal variation of giant<br />

clams, tube worms and mussels at <strong>de</strong>ep-sea hydrothermal vents. Bulletin of the<br />

Biological Society of Washington 6, 411-428.<br />

Hessler RR, Smithey WM (1983) The distribution and community structure of megafauna<br />

at the Galapagos Rift hydrothermal vents. In hydrothermal processes at seafloor<br />

spreading centres. Rona PA, Bostrom K, Laubier L, Smith KL Jr (eds), Plenum, New<br />

York.<br />

Hey J, Nielsen R (2004) Multilocus methods for estimating population sizes, migration<br />

rates and divergence time, with applications to the divergence of Drosophila<br />

pseudoobscura and D. persimilis. Genetics 167, 747-760.<br />

Hey J, Nielsen R (2007) Integration within the Felsenstein equation for improved Markov<br />

chain Monte Carlo methods in population genetics. Proceedings of the National<br />

Aca<strong>de</strong>my of Sciences of the United States of America 104, 2785-2790.<br />

Hey R, Baker E, Bohnenstiehl D, et al. (2004) Tectonic/volcanic segmentation and<br />

controls on hydrothermal venting along Earth's fastest seafloor spreading system, EPR<br />

27 <strong>de</strong>grees-32 <strong>de</strong>grees S. Geochemistry Geophysics Geosystems 5.<br />

Hickerson MJ, Dolman G, Moritz C (2006a) Comparative phylogeographic summary<br />

statistics for testing simultaneous vicariance. Molecular Ecology 15, 209-223.<br />

Hickerson MJ, Meyer CP (2008) Testing comparative phylogeographic mo<strong>de</strong>ls of marine<br />

vicariance and dispersal using a hierarchical Bayesian approach. Bmc Evolutionary<br />

Biology 8.<br />

Hickerson MJ, Stahl E, Takebayashi N (2007) msBayes: Pipeline for testing comparative<br />

phylogeographic histories using hierarchical approximate Bayesian computation.<br />

Bmc Bioinformatics 8.<br />

Hickerson MJ, Stahl EA, Lessios HA (2006b) Test for simultaneous divergence using<br />

approximate Bayesian computation. Evolution 60, 2435-2453.<br />

Ho SYW (2009) An examination of phylogenetic mo<strong>de</strong>ls of substitution rate variation<br />

among lineages. Biology letters. doi:10.1098/rsbl.2008.0729<br />

Hoelzer GA (1997) Inferring phylogenies from mtDNA variation: Mitochondrial-gene<br />

trees versus nuclear-gene trees revisited. Evolution 51, 622-626.<br />

Howell N, Bogolin Smejkal C, Mackey DA, Chinnery PF, Turnbull DM, Herrnstadt C<br />

(2003) The pedigree rate of sequence divergence in the human mitochondrial<br />

genome: there is a différence between phylogenetic and pedigree rates. The<br />

American Journal of Human Genetics 72, 659-670.<br />

Hudson RR (1983) Testing the constant-rate neutral allele mo<strong>de</strong>l with protein sequence<br />

data. Evolution 37, 203-217.<br />

Hudson RR, Kaplan NL (1985) Statistical Properties of the Number of Recombination<br />

Events in the History of a Sample of DNA-Sequences. Genetics 111, 147-164.<br />

308


Références Bibliographiques<br />

Hudson RR, Kreitman M, Agua<strong>de</strong> M (1987) A Test of Neutral Molecular Evolution Based<br />

on Nucleoti<strong>de</strong> Data. Genetics 116, 153-159.<br />

Hudson RR, Slatkin M (1991) Pairwise comparisons of mitochondrial-DNA sequences in<br />

stable and exponentially growing populations. Genetics 129, 555-562.<br />

Hudson RR, Slatkin M, Maddison WP (1992) Estimation of Levels of Gene Flow from<br />

DNA-Sequence Data. Genetics 132, 583-589.<br />

Hudson RR, Turelli M (2003) Stochasticity overrules the "three-times rule": Genetic drift,<br />

genetic draft, and coalescence times for nuclear loci versus mitochondrial DNA.<br />

Evolution 57, 182-190.<br />

Hurtado L, Lutz R, Vrijenhoek R (2004) Distinct patterns of genetic differentiation among<br />

annelids of eastern Pacific hydrothermal vents. Molecular Ecology 13, 2603-2615.<br />

Hurtado LA, Mateos M, Lutz RA, Vrijenhoek RC (2002) Molecular evi<strong>de</strong>nce for multiple<br />

species of oasisia (annelida: Siboglinidae) at eastern pacific hydrothermal vents.<br />

Cahier <strong>de</strong> Biologie Marine, 43, 377-380.<br />

Hurtado LA, Mateos M, Lutz RA, Vrijenhoek RC (2003) Coupling of bacterial<br />

endosymbiont and host mitochondrial genomes in the hydrothermal vent clam<br />

Calyptogena magnifica. Applied and Environmental Microbiology 69, 2058-2064.<br />

I<br />

Irwin DE (2002) Phylogeographic breaks without geographic barriers to gene flow.<br />

Evolution 56, 2383-2394.<br />

J<br />

Jennings WB, Edwards SV (2005) Speciational history of Australian grass finches<br />

(Poephila) inferred from thirty gene trees. Evolution 59, 2033-2047.<br />

Johannesen J, Johannesen B, Griebeler EM, et al. (2006) Distortion of symmetrical<br />

introgression in a hybrid zone: evi<strong>de</strong>nce for locus-specific selection and unidirectional<br />

range expansion. Journal of Evolutionary Biology 19, 705-716.<br />

Johnson SB, Waren A, Vrijenhoek RC (2008) DNA barcoding of Lepetodrilus limpets<br />

reveals cryptic species. Journal of Shellfish Research 27, 43-51.<br />

Johnson SB, Young CR, Jones WJ, Waren A, Vrijenhoek RC (2006) Migration, isolation,<br />

and speciation of hydrothermal vent limpets (Gastropoda; Lepetodrilidae) across the<br />

Blanco transform fault Biological Bulletin 210, 140-157.<br />

Jollivet D & Party TSS (2005) The biospeedo cruise: A new survey of hydrothermal vents<br />

along the south east pacific rise from 7°24s to 21°33s. InterRidge News, 13, 20-26.<br />

Jollivet D (1993) Distribution et évolution <strong>de</strong> la faune associée aux sources hydrothermales<br />

profon<strong>de</strong>s à 13°N sur la dorsale du Pacifique oriental : le cas particulier <strong>de</strong>s polychètes<br />

Alvinellidae. Volume 1. Thèse <strong>de</strong> doctorat, Université <strong>de</strong> Bretagne Occi<strong>de</strong>ntale, Brest.<br />

Jollivet D (1996) Specific and genetic diversity at <strong>de</strong>ep-sea hydrothermal vents: An<br />

overview. Biodiversity and Conservation 5, 1619-1653.<br />

Jollivet D, Chevaldonne P, Planque B (1999) Hydrothermal-vent alvinellid polychaete<br />

dispersal in the eastern Pacific. 2. A metapopulation mo<strong>de</strong>l based on habitat shifts.<br />

Evolution 53, 1128-1142.<br />

Jollivet D, Desbruyeres D, Bonhomme F, Moraga D (1995) Genetic Differentiation of<br />

Deep-Sea Hydrothermal Vent Alvinellid Populations (Annelida, Polychaeta) Along<br />

the East Pacific Rise. Heredity 74, 376-391.<br />

309


Références Bibliographiques<br />

Jollivet D, Empis A, Baker MC, et al. (2000) Reproductive biology, sexual dimorphism,<br />

and population structure of the <strong>de</strong>ep sea hydrothermal vent scale-worm,<br />

Branchipolynoe seepensis (Polychaeta : Polynoidae). Journal of the Marine Biological<br />

Association of the United Kingdom 80, 55-68.<br />

Jolly MT, Viard F, Gentil F, Thiebaut E, Jollivet D (2006) Comparative phylogeography of<br />

two coastal polychaete tubeworms in the Northeast Atlantic supports shared history<br />

and vicariant events. Molecular Ecology 15, 1841-1855.<br />

Jukes TH, Cantor CR (1969) Evolution of protein molecules. In: Mammalian Protein<br />

Metabolism (ed. Munro HN), pp 21-123. Aca<strong>de</strong>mic Press, New York.<br />

Juniper SK, Tunnicliffe V (1997) Crustal accretion and the hot vent ecosystem.<br />

Philosophical Transactions of the Royal Society of London. Series A, 355, 459-474.<br />

K<br />

Karl SA, Schutz S, Desbruyeres D, Lutz R, Vrijenhoek RC (1996) Molecular analysis of<br />

gene flow in the hydrothermal vent clam (Calyptogena magnifica). Molecular Marine<br />

Biology and Biotechnology 5, 193-202.<br />

Keenan CP (1991) Phylogeny of Australian Species of Flatheads (Teleostei,<br />

Platycephalidae) as Determined by Allozyme Electrophoresis. Journal of Fish Biology<br />

39, 237-249.<br />

Kelly NE, Metaxas A (2007) Influence of habitat on the reproductive biology of the <strong>de</strong>epsea<br />

hydrothermal vent limpet Lepetodrilus fucensis (Vetigastropoda : Mollusca) from<br />

the Northeast Pacific. Marine Biology 151, 649-662.<br />

Kervévan (1991) Modélisation tridimentionnelle <strong>de</strong> la convection hydrothermale en zone<br />

<strong>de</strong> rift océanique. Ph.D. dissertation, Université <strong>de</strong> Bretagne Occi<strong>de</strong>ntale, Brest,<br />

France.<br />

Kimura M (1953) "Stepping stone" mo<strong>de</strong>l of population. Annual Report Nationnal Institute<br />

of Genetics. Japan 3, 62-63.<br />

Kimura M (1980) A simple method for estimating evolutionary rates of base substitutions<br />

through comparative studies of nucleoti<strong>de</strong> sequences. Journal of Molecular Evolution<br />

16, 111-120.<br />

Kimura M, Weiss GH (1964) Stepping Stone Mo<strong>de</strong>l of Population Structure + Decrease of<br />

Genetic Correlation with Distance. Genetics 49, 561<br />

Kingman JFC (1982) The coalescent. Stochastic Process Appl 13.<br />

Kinlan BP, Gaines SD (2003) Propagule dispersal in marine and terrestrial environments:<br />

A community perspective. Ecology 84, 2007-2020.<br />

Knowlton N (1993) Sibling species in the sea. Ann. Rev. Ecol. Syst., 24, 189-216.<br />

Knowlton N (2000) Molecular genetic analyses of species boundaries in the sea.<br />

Hydrobiologia 420, 73-90.<br />

Knowlton N, Weigt LA (1998) New dates and new rates for divergence across the Isthmus<br />

of Panama. Proceedings of the Royal Society of London Series B-Biological<br />

Sciences 265, 2257-2263.<br />

Kojima S, Fujikura K, Okutani T (2004) Multiple trans-Pacific migrations of <strong>de</strong>ep-sea<br />

vent/seep-en<strong>de</strong>mic bivalves in the family Vesicomyidae. Molecular Phylogenetics and<br />

Evolution 32, 396-406.<br />

Kruuk LEB, Baird SJE, Gale KS, Barton NH (1999) A comparison of multilocus clines<br />

maintained by environmental adaptation or by selection against hybrids. Genetics 153,<br />

1959-1971.<br />

310


Références Bibliographiques<br />

Kuhner MK, Yamato J, Felsenstein J (1998) Maximum likelihood estimation of population<br />

growth rates based on the coalescent. Genetics 149, 429-434.<br />

Kuhner MK, Yamato J, Felsenstein J (1998) Maximum likelihood estimation of population<br />

growth rates based on the coalescent. Genetics 149, 429-434.<br />

Kumar S, Balczarek KA, Lai Z-C (1996) Evolution of the hedgehog gene family. Genetics<br />

142, 965-972.<br />

Kuo CH, Avise J (2005) Phylogeographic breaks in low-dispersal species: the emergence<br />

of concordance across gene trees. Genetica 124, 179-186.<br />

Kureth CL, Rea DK (1981) Large-Scale Oblique Features in an Active Transform-Fault,<br />

the Wilkes Fracture-Zone near 9-Degrees-S on the East Pacific Rise. Marine<br />

Geophysical Researches 5, 119-137.<br />

L<br />

Larson RL, Searle RC, Kleinrock MC, et al. (1992) Roller-bearing tectonic evolution of<br />

the Juan Fernan<strong>de</strong>z microplate. Nature 356 571-575.<br />

Laubier L (1986) Des oasis au fond <strong>de</strong>s mers. Monaco: Le Rocher (Sciences et<br />

Découvertes).<br />

Lavery L, Moritz C, Fiel<strong>de</strong>r DR (1995) Changing patterns of population structure and gene<br />

flow at different spatial scales in Birgus latro (the coconut crab). Heredity 74, 531-<br />

541.<br />

Le Bris N, Govenar BW, Le Gall C, Fisher CR (2006) Variability of physico-chemical<br />

conditions in 9°50'n epr diffuse flow vent habitats. Marine Chemistry, 98, 167-182.<br />

Le Pennec M (1988) Alimentation et reproduction d’un Mytilidae <strong>de</strong>s sources<br />

hydrothermales profon<strong>de</strong>s du Pacifique oriental. Oceanologica Acta 8, 181-190.<br />

Le Pennec M, Beninger PG (1997) Ultrastructural characteristics of spermatogenesis in<br />

three species of <strong>de</strong>ep-sea hydrothermal vent mytilids. Canadian Journal of<br />

Zoology-Revue Canadienne De Zoologie 75, 308-316.<br />

Le Pennec M, Hily A, Lucas A (1984) Structures gonadiques particulières d’un Mytilidae<br />

profound <strong>de</strong>s sources hydrothermales du Pacifique oriental. Comptes Rendus <strong>de</strong><br />

l’Académie <strong>de</strong>s Sciences <strong>de</strong> Paris, Série III 299, 725-730.<br />

Legendre P, Legendre L (1998) Numerical ecology. Elsevier.<br />

Lemaire C, Versini JJ, Bonhomme F (2005) Maintenance of genetic differentiation across<br />

a transition zone in the sea: discordance between nuclear and cytoplasmic markers.<br />

Journal of Evolutionary Biology 18, 70-80.<br />

Lessa EP, Cook JA, Patton JL (2003) Genetic footprints of <strong>de</strong>mographic expansion in<br />

North America, but not Amazonia, during the Late Quaternary Proceedings of the<br />

National Aca<strong>de</strong>my of Sciences of the United States of America 100, 10331-10334.<br />

Levin LA, Etter RJ, Rex MA, Gooday J, Smith CR, Pineda J, Stuart CT, Hessler RR,<br />

Pawson D (2001) Environmental influences on regional <strong>de</strong>ep-sea species diversity.<br />

Annual Review of Ecology and Systematics 32, 51-93.<br />

Levin R (1970) Extinction. Lectures on Mathematics in the Life Sciences 2, 75-107.<br />

Lewontin RC, Krakauer J (1973) Distribution of Gene Frequency as a Test of Theory of<br />

Selective Neutrality of Polymorphisms. Genetics 74, 175-195.<br />

Li W-H, Graur D (1991) Fundamentals of molecular evolution. Sinauer Associates, Inc.<br />

Sun<strong>de</strong>rland, Massachusetts.<br />

Little CTS, Vrijenhoek RC (2003) Are hydrothermal vent animals living fossils? Trends in<br />

ecology and evolution, 18, 582-588.<br />

311


Références Bibliographiques<br />

Liu HP, Mitton JB, Wu SK (1996) Paternal mitochondrial DNA differentiation far exceeds<br />

maternal mitochondrial DNA and allozyme differentiation in the freshwater mussel,<br />

Anodonta grandis grandis. Evolution 50, 952-957.<br />

Liu LL, Foltz DW, Stickle WB (1991) Genetic Population-Structure of the Southern<br />

Oyster Drill Stramonita (= Thais) Haemostoma. Marine Biology 111, 71-79.<br />

Lonsdale P (1977) Clustering of suspension-feeding macorbenthos near abyssal<br />

hydrothermal vents at oceanic spreading centres. Deep-sea Research 24, 857-863.<br />

Lorion J, Duperron S, Gros O, Cruaud C, Samadi S (2009) Several <strong>de</strong>ep-sea mussels and<br />

their associated symbionts are able to live both on wood and on whale falls.<br />

Proceedings of the Royal Society B-Biological Sciences 276, 177-185.<br />

Lupton J (1998) Hydrothermal helium plumes in the Pacific Ocean. Journal of<br />

Geophysical Research-Oceans 103, 15853-15868.<br />

Lupton JE, Craig H (1981) A Major He-3 Source at 15-Degrees-S on the East Pacific Rise.<br />

Science 214, 13-18.<br />

Lutz RA (1988) Dispersal of organisms at <strong>de</strong>ep-sea hydrothermal vents : A review.<br />

Oceanologica Acta. Hydrothermalism, Biology and Ecology Symposium, Paris, 4-7<br />

November 1985, Proceedings 23-29.<br />

Lutz RA, Fritz LW, Cerrato RM (1988) A Comparison of Bivalve (Calyptogena-<br />

Magnifica) Growth at 2 Deep-Sea Hydrothermal Vents in the Eastern Pacific. Deep-<br />

Sea Research Part a-Oceanographic Research Papers 35, 1793<br />

Lutz RA, Jablonski D, Rhoads DC, Turner RD (1980) Larval Dispersal of a Deep-Sea<br />

Hydrothermal Vent Bivalve from the Galapagos Rift. Marine Biology 57, 127-133.<br />

Lutz RA, Shank TM, Fornari DJ, et al (1994) Rapid growth at <strong>de</strong>ep-sea vents. Nature 371,<br />

663-664.<br />

Lyrholm T, Leimar O, Johanneson B, Gyllensten U (1999) Sex-biased dispersal in sperm<br />

whales: contrasting mitochondrial and nuclear genetic structure of global populations.<br />

Proceeding of the Royal Society of London Series B 266: 347-354.<br />

M<br />

Macdonald KC, Scheirer DS, Carbotte SM (1991) Mid-Ocean Ridges: Discontinuities,<br />

Segments and Giant Cracks. Science 253, 986-994.<br />

Maddison WP (1997) Gene trees in species trees. Systematic Biology 46, 523-536.<br />

Mag<strong>de</strong> LS, Sparks DW (1997) Three-dimensional mantle upwelling, melt generation and<br />

melt migration beneath segmented slow-spreading ridges. Journal of Geophysical<br />

Research 102, 20571–20583.<br />

Maggs CA, Castilho R, Foltz D, et al. (2008) Evaluating signatures of glacial refugia for<br />

north atlantic benthic marine taxa. Ecology 89, 108-122.<br />

Malécot G (1950) Quelques schémas probabilistes sur la variabilité <strong>de</strong>s populations<br />

naturelles. Annales <strong>de</strong> l’Université <strong>de</strong> Lyon Sciences 13, 37-60.<br />

Mammerickx J, Herron E, Dorman L (1980) Evi<strong>de</strong>nce for two fossil spreading ridges in<br />

the southeast Pacific. Geological Society of America Bulletin 91, 263-271.<br />

Mammerickx J, Klitgord KD (1982) Northern East Pacific Rise - Evolution from 25 My<br />

Bp to the Present. Journal of Geophysical Research 87, 6751-6759.<br />

Mantel N (1967) The <strong>de</strong>tection of disease clustering and a generalized regression<br />

approach. Cancer Research 27, 209-220.<br />

Marcus J, Tunnicliffe V, Butterfield DA (2009) Post-eruption succession of macrofaunal<br />

communities at diffuse flow hydrothermal vents on Axial Volcano, Juan <strong>de</strong> Fuca<br />

312


Références Bibliographiques<br />

Ridge, Northeast Pacific. Deep-Sea Research Part Ii-Topical Studies in Oceanography<br />

56, 1586-1598.<br />

Marsh AG, Mullineaux LS, Young CM, Manahan DT (2001) Larval dispersal potential of<br />

the tubeworm Riftia pachyptila at <strong>de</strong>ep-sea hydrothermal vents. Nature 411, 77-80.<br />

Martinsen GD, Whitham TG, Turek RJ, Keim P (2001) Hybrid populations selectively<br />

filter gene introgression between species. Evolution 55, 1325-1335.<br />

Matabos M, Le Bris N, Pendlebury S, Thiebaut E (2008a) Role of physico-chemical<br />

environment on gastropod assemblages at hydrothermal vents on the East Pacific Rise<br />

(13 <strong>de</strong>grees N/EPR). Journal of the Marine Biological Association of the United<br />

Kingdom 88, 995-1008.<br />

Matabos M, Thiebaut E, Le Guen D, et al. (2008b) Geographic clines and stepping-stone<br />

patterns <strong>de</strong>tected along the EPR in the vetigastropod Lepetodrilus elevatus reflects<br />

species crypticism. Marine Biology 153, 545-563.<br />

McDonald JH, Kreitman M (1991b) Adaptive Protein Evolution at the Adh Locus in<br />

Drosophila. Nature 351, 652-654.<br />

McDonald JH, Kreitman M (1991a) Neutral Mutation Hypothesis Test - Reply. Nature<br />

354, 116-116.<br />

McLean JH (1988) New Archaeogastropod limpets from hydrothermal vents; superfamily<br />

Lepetodrilacea I. systematic <strong>de</strong>scriptions. Philosophical Transactions of the Royal<br />

Society of London 319, 1-32.<br />

Metaxas A (2004) Spatial and temporal patterns in larval supply at hydrothermal vents in<br />

the northeast Pacific Ocean. Limnology and Oceanography 49, 1949-1956.<br />

Metropolis N, Rosenbluth AW, Rosenbluth MN, Teller AH, Teller E (1953) Equation of<br />

State Calculations by Fast Computing Machines. Journal of Chemical Physics 21,<br />

1087-1092.<br />

Micheli F, Peterson CH, Mullineaux LS, Fisher CR, Mills SW, Sancho G, Johnson GA,<br />

Lenihan HS (2002) Predation structures communties at <strong>de</strong>ep-sea hydrothermal<br />

vents. Ecological monographs, 72, 365-382.<br />

Mills SW, Mullineaux LS, Tyler PA (2007) Habitat associations in gastropod species at<br />

east Pacific rise hydrothermal vents (9 <strong>de</strong>grees 50 ' N). Biological Bulletin 212, 185-<br />

194.<br />

Monge-Najera J (2008) Ecological biogeography: A review with emphasis on conservation<br />

and the neutral mo<strong>de</strong>l. Gayana, 72, 102-112.<br />

Moore WS (1995) Inferring Phylogenies from Mtdna Variation - Mitochondrial-Gene<br />

Trees Versus Nuclear-Gene Trees. Evolution 49, 718-726.<br />

Moritz C, Faith DP (1998) Comparative phylogeography and the i<strong>de</strong>ntification of<br />

genetically divergent areas for conservation. Molecular Ecology 7, 419-429.<br />

Morjan CL, Rieseberg LH (2004) How species evolve collectively: implications of gene<br />

flow and selection for the spread of advantageous alleles. Molecular Ecology 13,<br />

1341-1356.<br />

Muirhead CA (2001) Consequences of population structure on genes un<strong>de</strong>r balancing<br />

selection. Evolution 55, 1532-1541.<br />

Mullineaux LS, Fisher CR, Peterson CH, Schaeffer SW (2000) Tubeworm succession at<br />

hydrothermal vents: Use of biogenic cues to reduce habitat selection error?<br />

Oecologia, 123, 275-284Mullineaux, L.S., Mills, S.W. & Goldman, E. (1998)<br />

Recruitment variation during a pilot colonization study of hydrothermal vents<br />

(9°50'n, east pacific rise). Deep Sea Research II, 45, 441-464.<br />

Mullineaux LS, Kim SL, Pooley A, Lutz RA (1996) I<strong>de</strong>ntification of archaeogastropod<br />

larvae from a hydrothermal vent community. Marine Biology 124, 551-560.<br />

313


Références Bibliographiques<br />

Mullineaux LS, Mills SW, Goldman E (1998) Recruitment variation during a pilot<br />

colonization study of hydrothermal vents (9 <strong>de</strong>grees 50 ' N, East Pacific Rise). Deep-<br />

Sea Research Part Ii-Topical Studies in Oceanography 45, 441-464.<br />

Mullineaux LS, Mills SW, Sweetman AK, Beaudreau AH, Metaxas A, Hunt HL (2005)<br />

Vertical, lateral and temporal structure in larval distributions at hydrothermal vents.<br />

Marine Ecology Progress Series 293, 1-16.<br />

Mullineaux LS, Peterson CH, Micheli F, Mills SW (2003) Successional mechanism varies<br />

along a gradient in hydrothermal fluid flux at <strong>de</strong>ep-sea vents. Ecological<br />

monographs, 73, 523-542.<br />

Mullineaux LS, Wiebe PH, Baker ET (1995) Larvae of benthic invertebrates in<br />

hydrothermal vent plumes over Juan-De-Fuca Ridge. Marine Biology 122, 585-596.<br />

Myers AA, Giller PS (1988) Analytical Biogeography : an integrated approach to the study<br />

of animal and plant distributions. Chapman & Hall, London.<br />

N<br />

Naar DF, Hey RN (1989) Speed Limit for Oceanic Transform Faults. Geology 17, 420-<br />

422.<br />

Naar DF, Hey RN (1991) Tectonic evolution of the Easter microplate. Journal of<br />

Geophysical Research 96, 7961–7993.<br />

Nagylaki T (1975) Conditions for Existence of Clines. Genetics 80, 595-615.<br />

Nei M, Kumar S (2001) Molecular evolution and phylogenetics. Oxford University Press,<br />

UK.<br />

Neigel JE, Avise JC (1986) Phylogenetic relationships of mitochondrial DNA un<strong>de</strong>r<br />

various <strong>de</strong>mographic mo<strong>de</strong>ls of speciation. Karling S & Nevo E (eds). In:<br />

Evolutionary Process and Theory. Aca<strong>de</strong>mic Press, New York.<br />

Neigel JE, Avise JC (1993) Application of a Random-Walk Mo<strong>de</strong>l to Geographic<br />

Distributions of Animal Mitochondrial-DNA Variation. Genetics 135, 1209-1220.<br />

Nielsen R, Wakeley J (2001) Distinguishing migration from isolation: A Markov chain<br />

Monte Carlo approach. Genetics 158, 885-896.<br />

Nosil P, Harmon LJ, Seehausen O (2009) Ecological explanations for (incomplete)<br />

speciation. Trends in Ecology & Evolution 24, 145-156.<br />

O<br />

OhUigin C, Sato A, Klein J (1997) Evi<strong>de</strong>nce for convergent evolution of A and B blood<br />

group antigens in primates. Human Genetics 101, 141-148.<br />

Olu-Le Roy L, von Cosel R, Hour<strong>de</strong>z S, Carney SL, Jollivet D (2007) Amphi-Atlantic<br />

cold-seep Bathymodiolus species complexes across the equatorial belt. Deep-Sea<br />

Research I 54, 1890-1911.<br />

O'Mullan GD, Maas PAY, Lutz RA, Vrijenhoek RC (2001) A hybrid zone between<br />

hydrothermal vent mussels (Bivalvia : Mytilidae) from the Mid-Atlantic Ridge.<br />

Molecular Ecology 10, 2819-2831.<br />

P<br />

314


Références Bibliographiques<br />

Page RDM, Holmes EC (1998). Measuring genetic change, In Molecular evolution : a<br />

phylogenetic approach. Blackwell Science, London, 153.<br />

Palumbi SR (1994) Genetic-Divergence, Reproductive Isolation, and Marine Speciation.<br />

Annual Review of Ecology and Systematics 25, 547-572.<br />

Palumbi SR (1996) What can molecular genetics contribute to marine biogeography?<br />

Journal of Experimental Marine Biology and Ecology, 203, 75-92.<br />

Palumbi SR (2003) Population genetic, <strong>de</strong>mographic connectivity, and the <strong>de</strong>sign of<br />

marine reserves. Ecological Applications 13, 146-158.<br />

Pamilo P, Nei M (1988) Relationships between Gene Trees and Species Trees. Molecular<br />

Biology and Evolution 5, 568-583.<br />

Pamilo P, Viljakainen L, Vihavainen A (2007) Exceptionally high <strong>de</strong>nsity of NUMTs in<br />

the honeybee genome. Molecular Biology and Evolution 24, 1340-1346.<br />

Pannell JR, Charlesworth B (2000) Effects of metapopulation process on measures of<br />

genetic diversity. Philosophical Transactions of the Royal Society of London 355,<br />

1851-1864.<br />

Passamonti M, Scali V (2001) Gen<strong>de</strong>r-associated mitochondrial DNA heteroplasmy in the<br />

venerid clam Tapes philippinarum (Mollusca Bivalvia). Current Genetics 39, 117-124.<br />

Pasteur N, Pasteur G, Bonhomme F, Catalan J, Britton-Davidian J (1987) Manuel<br />

technique <strong>de</strong> génétique par électrophorèse <strong>de</strong> protéines. Paris: Lavoisier.<br />

Pechenik JA (1999) On the advantages ans disadvantages of larval stages in benthic marine<br />

invertebrate life cycles. Marine Ecology Progress Series 177, 269-297.<br />

Peek AS, Gustafson RG, Lutz RA, Vrijenhoek RC (1997) Evolutionary relationships of<br />

<strong>de</strong>ep-sea hydrothermal vent and cold-water seep clams (bivalvia: Vesicomyidae):<br />

Results from the mitochondrial cytochrome oxidase subunit i. Marine Biology, 130,<br />

151-161.<br />

Peters JL, Zhuravlev Y, Fefelov I, Logie A, Omland KE (2007) Nuclear loci and<br />

coalescent methods support ancient hybridization as cause of mitochondrial<br />

paraphyly between gadwall and falcated duck (Anas spp.). Evolution 61, 1992-<br />

2006.<br />

Piálek J, Barton NH (1997) The spread of an advantageous allele across a barrier: The<br />

effects of random drift and selection against heterozygotes. Genetics 145, 493-504.<br />

Piccino P, Viard F, Sarradin PM, et al (2004) Thermal selection of PGM allozymes in<br />

newly foun<strong>de</strong>d populations of the thermotolerant vent polychaete Alvinella<br />

pompejana. Proceedings of the Royal Society of London Series B-Biological Sciences<br />

271, 2351-2359.<br />

Plouviez S (2006) Divergence et polymorphisme nucléotidiques chez les Lepetodrilidae,<br />

complexe d’espèces hydrothermales partiellement sympatriques inféodées à <strong>de</strong>s<br />

microhabitats différents. Rapport <strong>de</strong> Master 2. Université Montpellier 2.<br />

Plouviez S, Daguin C, Hour<strong>de</strong>z S, Jollivet D (2008) Juvenile and adult scale-worms,<br />

Branchipolynoe seepensis, in Lucky Strike hydrothermal vent mussels are<br />

genetically unrelated. Aquatic Biology 3, 79-87.<br />

Plouviez S, Shank TM, Faure B, Daguin-Thiébaut C, Viard F, Lallier FH, Jollivet D<br />

(2009) Comparative phylogeography among hydrothermal vent species along the East<br />

Pacific Rise reveals vicariant processes and population expansion in the South.<br />

Molecular Ecology 18: 3903-3917.<br />

Pluzhnikov A, Donnelly P (1996) Optimal sequencing strategies for surveying molecular<br />

genetic diversity. Genetics 144, 1247-1262.<br />

Pogson GH, Mesa KA, Boutilier RG (1995) Genetic Population-Structure and Gene Flow<br />

in the Atlantic Cod Gadus-Morhua - a Comparison of Allozyme and Nuclear Rflp<br />

Loci. Genetics 139, 375-385.<br />

315


Références Bibliographiques<br />

Posada D, Crandall KA (2001) Intraspecific gene genealogies: trees grafting into networks.<br />

Trends in Ecology & Evolution 16, 37-45.<br />

Pradillon F, Gaill F (2003) Oogenesis characteristics in the hydrothermal vent polychaete<br />

Alvinella pompejana. Invertebrate Reproduction & Development 43, 223-235.<br />

Pradillon F, Le Bris N, Shillito B, Young CM, Gaill F (2005) Influence of environmental<br />

conditions on early <strong>de</strong>velopment of the hydrothermal vent polychaete Alvinella<br />

pompejana. Journal of Experimental Biology 208, 1551-1561.<br />

Pradillon F, Shillito B, Young CM, Gaill F (2001) Deep-seaecology - Developmental<br />

arrest in vent worm embryos. Nature 413, 698-699.<br />

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using<br />

multilocus genotype data. Genetics 155, 945-959.<br />

Przeworski M, Wall JD, Andolfatto P (2001) Recombination and the frequency spectrum<br />

in Drosophila melanogaster and Drosophila simulans. Molecular Biology and<br />

Evolution 18, 291-298.<br />

R<br />

Rambaut A, Bromham L (1998) Estimating divergence dates from molecular sequences.<br />

Molecular Biology and Evolution 15, 442-448.<br />

Rand DM (2001) The units of selection on mitochondrial DNA. Annual Review of Ecology<br />

and Systematics 32, 415-448.<br />

Raulfs EC, Macko SA, Van Dover CL (2004) Tissue and symbiont condition of mussels<br />

(Bathymodiolus thermophilus) exposed to varying levels of hydrothermal activity.<br />

Journal of the Marine Biological Association of the United Kingdom 84, 229-234.<br />

Reid JL (1997) On the total geostrophic circulation of the South-Pacific ocean: flow<br />

patterns, tracers, and transports. Progress in Oceanography 39, 263-352.<br />

Reid RA (1980) Selfish DNA in "petite" mutants. Nature 285, 620.<br />

Renard V, Hekinian R, Francheteau J, Ballard RD, Backer H (1985) Submersible<br />

Observations at the Axis of the Ultra-Fast-Spreading East Pacific Rise (17-Degrees-'<br />

to 21-Degrees-30's). Earth and Planetary Science Letters 75, 339-353.<br />

Rieseberg LH, Whitton J, Gardner K (1999) Hybrid zones and the genetic architecture of a<br />

barrier to gene flow between two sunflower species. Genetics 152: 713-727.<br />

Rodríguez F, Olivier JL, Marín A, Medina JR (1990) The general stochastic mo<strong>de</strong>l of<br />

nucleoti<strong>de</strong> substitution. Journal of Theoretical Biology 142, 485-501.<br />

Rolan-alvarez E, Zapata C, Alvarez G (1995) Distinct genetic subdivision in sympatric and<br />

sibling species of the genus Littorina (Gastropoda, Littorinidae). Heredity 74, 1-9.<br />

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA<br />

polymorphism analyses by the coalescent and other methods. Bioinformatics 19, 2496-<br />

2497.<br />

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA<br />

polymorphism analyses by the coalescent and other methods. Bioinformatics 19,<br />

2496-2497.<br />

Rusby RI, Searle RC (1995) A History of the Easter Microplate, 5.25-Ma to Present.<br />

Journal of Geophysical Research-Solid Earth 100, 12617-12640.<br />

S<br />

316


Références Bibliographiques<br />

Sadosky F, Thiebaut E, Jollivet D, Shillito B (2002) Recruitment and population structure<br />

of the vetigastropod Lepetodrilus elevatus at 13 <strong>de</strong>grees N hydrothermal vent sites<br />

on East Pacific Rise. Cahiers <strong>de</strong> Biologie Marine 43, 399-402.<br />

Samadi S, Quemere E, Lorion J, et al. (2007) Molecular phylogeny in mytilids supports<br />

the woo<strong>de</strong>n steps to <strong>de</strong>ep-sea vents hypothesis. Comptes Rendus Biologies 330,<br />

446-456.<br />

Sancho G, Fisher CR, Mills SW, Micheli F, Johnson GA, Lenihan HS, Peterson CH,<br />

Mullineaux LS (2005) Selective predation by the zoarcid fish thermarces cerberus<br />

at hydrothermal vents. Deep Sea Research I, 52, 837-844.<br />

Santiago E, Caballero A (2005) Variation after a selective sweep in a subdivi<strong>de</strong>d<br />

population. Genetics 169, 475-483.<br />

Sarrazin J, Juniper SK (1999) Biological characteristics of a hydrothermal edifice mosaic<br />

community. Marine Ecology Progress Series, 185, 1-19.<br />

Sarrazin J, Juniper SK, Massoth GJ, Legendre P (1999) Physical and chemical factors<br />

influencing species distributions on hydrothermal sulfi<strong>de</strong> edifices of the juan <strong>de</strong><br />

fuca ridge, northeast pacific. Marine Ecology Progress Series, 190, 89-112<br />

Sarrazin J, Robigou V, Juniper SK, Delaney JR (1997) Biological and geological dynamics<br />

over four years on a high-temperature sulfi<strong>de</strong> structure at the juan <strong>de</strong> fuca ridge<br />

hydrothermal observatory. Marine Ecology Progress Series, 153, 5-24.<br />

Sarrazin J, Walter C, Sarradin P-M, Brind'amour A, Desbruyères D, Briand P, Fabri M-C,<br />

Van Gaever S, Vanreusel A, Bachraty C, Thiebaut E (2006) Community structure<br />

and temperature dynamics within a mussel assemblage in the southern east pacific<br />

rise. Cahiers <strong>de</strong> Biologie Marine, 47, 483-490.<br />

Schierup MH, Charlesworth D, Vekemans X (2000a) The effect of hitch-hiking on genes<br />

linked to a balanced polymorphism in a subdivi<strong>de</strong>d population. Genetical Research<br />

76, 63-73.<br />

Schierup MH, Hein J (2000) Consequences of recombination on traditional phylogenetic<br />

analysis. Genetics 156, 879-891.<br />

Schierup MH, Vekemans X, Charlesworth D (2000b) The effect of subdivision on<br />

variation at multi-allelic loci un<strong>de</strong>r balancing selection. Genetical Research 76, 51-62.<br />

Schlager W (2008) Earth’s layers, their cycles and Earth system science. Austrian Journal<br />

of Earth Sciences 101, 4-16.<br />

Shank TM, Fornari DJ, Von Damm KL et al (1998) Temporal and spatial patterns of<br />

biological community <strong>de</strong>velopment at nascent <strong>de</strong>ep-sea hydrothermal vents (9°50’N,<br />

East Pacific Rise). Deep Sea Research Part II: Topical Studies in Oceanography 45,<br />

465-515.<br />

Shank TM, Halanych KM (2007) Toward a mechanistic un<strong>de</strong>rstanding of larval dispersal:<br />

insights from genomic fingerprinting of the <strong>de</strong>ep-sea hydrothermal vent tubeworm<br />

Riftia pachyptila. Marine Ecology 28, 25-35.<br />

Shriner D, Nickle DC, Jensen MA, Mullins JI (2003) Potential impact of recombination on<br />

sitewise approaches for <strong>de</strong>tecting positive natural selection. Genetical Research 81,<br />

115-121.<br />

Skibinski DOF, Gallagher C, Beynon CM (1994) Mitochondrial-DNA Inheritance. Nature<br />

368, 817-818.<br />

Slatkin M (1977) Gene flow and genetic drift in a species subject to frequent local<br />

extinction. Theoretical Population Biology 12, 253-262.<br />

Slatkin M (1977) Gene flow and genetic drift in a species subject to frequent local<br />

extinction. Theoretical Population Biology 12, 253-262.<br />

317


Références Bibliographiques<br />

Smith CI, Farrell BD (2005) Range expansions in the flightless longhorn cactus beetles,<br />

Moneilema gigas and Moneilema armatum, in response to Pleistocene climate<br />

changes. Molecular Ecology 14, 1025-1044.<br />

Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS (2006) Comparative<br />

phylogeography of unglaciated eastern North America. Molecular Ecology 15, 4261-<br />

4293.<br />

Spalding MD, Fox HE, Allen GR, Davidson N, Ferdaña ZA, Finlayson M, Halpern BS,<br />

Jorge MA, Lombana A, Lourie SA, Martin KD, Mcmanus E, Omolnar J, Recchi,<br />

CA, Robertson J (2007) Marine ecoregions of the world: A bioregionalization of<br />

coastal and shelf areas. BioScience, 57, 573-583.<br />

Stöhr S, Segonzac M (2006) Two new genera and species of ophiuroid (echino<strong>de</strong>rmata)<br />

from hydrothermal vents in the east pacific. Species diversity, 11, 7-32.<br />

Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology &<br />

Evolution 15, 199-203.<br />

T<br />

Tajima F (1989) Statistical-method for testing the neutral mutation hypothesis by DNA<br />

polymorphism. Genetics 123, 585-595.<br />

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular evolutionary genetics<br />

analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24, 1596-<br />

1599.<br />

Templeton AR (1981) Mechanisms of Speciation - a Population Genetic Approach. Annual<br />

Review of Ecology and Systematics 12, 23-48.<br />

Thiel M, Macaya EC, Acuna E, et al. (2007) The Humboldt Current System of northern<br />

and central Chile. Oceanography and Marine Biology, Vol 45 45, 195-344.<br />

Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of<br />

progressive multiple sequence alignment through sequence weighting, position<br />

specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673-<br />

4680.<br />

Thomson RE, Roth SE, Dymond J (1990) Near-inertial motion over a mid-ocean ridge:<br />

effects of topography and hydrothermal plumes. Journal of Geophysical Research 95:<br />

12961-12966.<br />

Thorrold SR, Zacherl DC, Levin LA (2007) Population Connectivity and Larval Dispersal<br />

Using Geochemical Signatures in Calcified Structures. Oceanography 20, 80-89.<br />

Ting CT, Tsaur SC, Wu CI (2000) The phylogeny of closely related species as revealed by<br />

the genealogy of a speciation gene, Odysseus. Proceedings of the National Aca<strong>de</strong>my<br />

of Sciences of the United States of America 97, 5313-5316.<br />

Trewick SA, Wallis GP, Morgan-Richards M (2000) Phylogeographical pattern correlates<br />

with Pliocene mountain building in the alpine scree weta (Orthoptera,<br />

Anostostomatidae). Molecular Ecology 9, 657-666.<br />

Tunnicliffe V (1988) Biogeography and Evolution of Hydrothermal-Vent Fauna in the<br />

Eastern Pacific-Ocean. Proceedings of the Royal Society of London Series B-<br />

Biological Sciences 233, 347-366.<br />

Tunnicliffe V (1991) The biology of hydrothermal vents: ecology and evolution.<br />

Oceanography and Marine Biology. An annual review 29, 319-407.<br />

Tunnicliffe V, Botros M, Deburgh ME, et al. (1986) Hydrothermal Vents of Explorer<br />

Ridge, Northeast Pacific. Deep-Sea Research Part a-Oceanographic Research Papers<br />

33, 401-412.<br />

318


Références Bibliographiques<br />

Tunnicliffe V, Embley RW, Hol<strong>de</strong>n JF, et al. (1997) Biological colonization of new<br />

hydrothermal vents following an eruption on Juan <strong>de</strong> Fuca Ridge. Deep-Sea Research<br />

Part I-Oceanographic Research Papers 44, 1627-1644.<br />

Tunnicliffe V, Fowler CMR (1996) Influence of sea-floor spreading on the global<br />

hydrothermal vent fauna. Nature, 379, 531-533.<br />

Tunnicliffe V, Juniper SK, <strong>de</strong> Burgh ME (1985) The hydrothermal vent community of<br />

Axial seamount, Juan <strong>de</strong> Fuca Ridge. Bulletin of Biological Socity of Washington 6,<br />

453-464.<br />

Tunnicliffe V, McArthur AG, McHugh D (1998) A biogeographical perspective of the<br />

<strong>de</strong>ep-sea hydrothermal vent fauna. Advances in Marine Biology, Vol 34 34, 353-442.<br />

Turnipseed M, Knick KE, Lipcius RN, Dreyer JC, Van Dover CL (2003) Diversity in<br />

mussel beds at <strong>de</strong>ep-sea hydrothermal vents and cold seep. Ecology letters, 6, 518-<br />

523.<br />

Tyler P, Young C (2003) Dispersal at hydrothermal vents: a summary of recent progress.<br />

Hydrobiologia 503, 9-19.<br />

Tyler PA, German CR, Ramirez-Llodra E, Van Dover CL (2003) Un<strong>de</strong>rstanding the<br />

biogeography of chemosynthetic ecosystems. Oceanologica Acta, 25, 227-241.<br />

Tyler PA, Pendlebury S, Mills SW, et al. (2008) Reproduction of gastropods from vents on<br />

the East Pacific Rise and the Mid-Atlantic Ridge. Journal of Shellfish Research 27,<br />

107-118.<br />

Tyler PA, Young CM (1999) Reproduction and dispersal at vents and cold seeps. Journal<br />

of the Marine Biological Association of the United Kingdom 79, 193-208.<br />

V<br />

Van Dover C, Trask JL (2000) Diversity at <strong>de</strong>ep-sea hydrothermal vent and intertidal<br />

mussel beds. Marine Ecology Progress Series, 195, 169-178.<br />

Van Dover CL (2002) Community structure of mussels beds at <strong>de</strong>ep-sea hydrothermal<br />

vents. Marine Ecology Progress Series, 230, 137-158.<br />

Van Dover CL (2003) Variation in community srtucture within hydrothermal vent mussel<br />

beds of the east pacific rise. Marine Ecology Progress Series, 253, 55-66.<br />

Van Dover CL, Doerries MB (2005) Community structure in mussel beds at logatchev<br />

hydrothermal vents and a comparison of macrofaunal species richness on slow- and<br />

fast-spreading mid-ocean ridges. Marine Ecology, 26, 110-120.<br />

Van Dover CL, German CR, Speer KG, Parson LM, Vrijenhoek RC (2002) Evolution and<br />

biogeography of <strong>de</strong>ep-sea vent and seep invertebrates. Science 295, 1253-1257.<br />

Van Dover CL, Humphris SE, Fornari D, et al. (2001a) Biogeography and ecological<br />

setting of Indian Ocean hydrothermal vents. Science 294, 818-823.<br />

Van Dover CL, Jenkins CD, Turnipseed M (2001b) Corralling of larvae in the <strong>de</strong>ep sea.<br />

Journal of the Marine Biological Association of the United Kingdom 81, 823-826.<br />

Van Dover CL, Trask JL, Gross J, Knowlton A (1999) Reproductive biology of free-living<br />

and commensal polynoid polychaetes at the Lucky Strike hydrothermal vent field<br />

(Mid-Atlantic Ridge). Marine Ecology-Progress Series 181, 201-214.<br />

Via S (2001) Sympatric speciation in animals: The ugly duckling grows up. Trends in<br />

Ecology & Evolution, 16, 381-390.<br />

Vrijenhoek RC (1997) Gene flow and genetic diversity in naturally fragmented<br />

metapopulations of <strong>de</strong>ep-sea hydrothermal vent animals. Journal of Heredity 88, 285-<br />

293.<br />

319


Références Bibliographiques<br />

W<br />

Wa<strong>de</strong> MJ, McCauley DE (1988) Extinction and recolonization: their effects on the genetic<br />

differentiation of local populations. Evolution 42, 995-1005.<br />

Wakeley J (2009) The structured coalescent. In : Coalescent theory : an introduction.<br />

Camp R (ed), Roberts and Company, 165-166.<br />

Wall JD, Andolfatto P, Przeworski M (2002) Testing mo<strong>de</strong>ls of selection and <strong>de</strong>mography<br />

in Drosophila simulans. Genetics 162, 203-216.<br />

Waren A, Bengtson S, Goffredi SK, Van Dover CL (2003) A hot-vent gastropod with iron<br />

sulfi<strong>de</strong> <strong>de</strong>rmal sclerites. Science 302, 1007-1007.<br />

Warén A, Bouchet P (2001) Gastropoda and Monoplacophora from hydrothermal vents<br />

and seeps; New taxa and records. Veliger 44, 116-231.<br />

Warwick RM, Clarke KR (2001) Practical measures of marine biodiversity based on<br />

relatedness of species. Oceanography and Marine Biology. An Annual Review, 39,<br />

207-231.<br />

Watremez P, Kervévan C (1990) Origin of Variations in Hydrothermal Activity - 1st<br />

Results of a Simple Numerical-Mo<strong>de</strong>l. Comptes Rendus De L Aca<strong>de</strong>mie Des Sciences<br />

Serie Ii 311, 153-158.<br />

Watterson GA (1975) Number of Segregating Sites in Genetic Mo<strong>de</strong>ls without<br />

Recombination. Theoretical Population Biology 7, 256-276.<br />

Weir B, Cockerham C (1984) Estimating F-statistics for the analysis of population<br />

structure. Evolution 38: 1358-1370.<br />

Whitlock MC, McCauley DE (1990) Some population genetic consequences of colony<br />

formation and extinction: genetic correlations within foundings groups. Evolution 44,<br />

1717-1724.<br />

Whitlock MC, McCauley DE (1999) Indirect measures of gene flow and migration: Fst not<br />

equal 1/(4Nm+1). Heredity 82, 117-125.<br />

Whittam TS, Nei M (1991) Neutral Mutation Hypothesis Test. Nature 354, 115-116.<br />

Wilke T, Davis GM (2000) Intraspecific mitochondrial sequence diversity in Hydrobia<br />

ulvae and Hydrobia ventrosa (Hydrobiidae : Rissooi<strong>de</strong>a : Gastropoda): Do their<br />

different life histories affect biogeographic patterns and gene flow? Biological<br />

Journal of the Linnean Society 70, 89-105.<br />

Wolstenholme DR, Clary DO (1985) Sequence evolution of Drosophila mitochondrial<br />

DNA. Genetics 109, 725-744.<br />

Won Y, Young CR, Lutz RA, Vrijenhoek RC (2003) Dispersal barriers and isolation<br />

among <strong>de</strong>ep-sea mussel populations (Mytilidae : Bathymodiolus) from eastern Pacific<br />

hydrothermal vents. Molecular Ecology 12, 169-184.<br />

Won YJ, Jones WJ, Vrijenhoek RC (2008) Absence of cospeciation between <strong>de</strong>ep-sea<br />

Mytilids and their thiotrophic endosymbionts. Journal of Shellfish Research 27,<br />

129-138.<br />

Wood AR, Turner G, Skibinski DOF, Beaumont AR (2003) Disruption of doubly<br />

uniparental inheritance of mitochondrial DNA in hybrid mussels (Mytilus edulis x M-<br />

galloprovincialis). Heredity 91, 354-360.<br />

Wood TE, Burke JM, Rieseberg LH (2005) Parallel genotypic adaptation: when evolution<br />

repeats itself. Genetica 123, 157-170.<br />

Wright S (1931) Evolution in men<strong>de</strong>lian populations. Genetics 16, 97-159.<br />

Wright S (1937) The distribution of gene frequencies in populations. Proceedings of the<br />

National Aca<strong>de</strong>my of Sciences of the United States of America 23, 307-320.<br />

320


Références Bibliographiques<br />

Wright S (1937) The distribution of gene frequencies in populations. Proceedings of the<br />

National Aca<strong>de</strong>my of Sciences of the United States of America 23, 307-320.<br />

Wright S (1943) Isolation by distance. Genetics 28, 114-138.<br />

Wright S (1977) Evolution and the genetics of populations. Experimental Results and<br />

Evolutionary Deductions. University of Chicago Press, Chicago, Vol. 3<br />

Y<br />

Young CR, Fujio S, Vrijenhoek RC (2008) Directional dispersal between mid-ocean<br />

ridges: <strong>de</strong>ep-ocean circulation and gene flow in Ridgeia piscesae. Molecular Ecology<br />

17, 1718-1731.<br />

Z<br />

Zacherl DC (2005) Spatial and temporal variation in statolith and protoconch trace<br />

elements as natural tags to track larval dispersal. Marine Ecology-Progress Series<br />

290, 145-163.<br />

Zal F, Desbruyères D, Join-Toulmond C (1994) Sexual dimorphism in Paralvinella<br />

grasslei, a polychaete annelid from <strong>de</strong>ep-sea hydrothermal vents. Comptes rendus<br />

<strong>de</strong> l’Académie <strong>de</strong>s Sciences Série III Sciences <strong>de</strong> la vie 317, 42-48.<br />

Zhang JZ, Kumar S (1997) Detection of convergent and parallel evolution at the amino<br />

acid sequence level. Molecular Biology and Evolution 14, 527-536.<br />

Zierenberg RA, Adams MWW, Arp AJ (2000) Life in extreme environments:<br />

Hydrothermal vents. Proceedings of the National Aca<strong>de</strong>my of Sciences of the United<br />

States of America 97, 12961-12962.<br />

Zink RM (2002) Methods in comparative phylogeography, and their application to<br />

studying evolution in the North American aridlands. Integrative and Comparative<br />

Biology 42, 953-959.<br />

Zink RM, Barrowclough GF (2008) Mitochondrial DNA un<strong>de</strong>r siege in avian<br />

phylogeography. Molecular Ecology 17, 2107-2121.<br />

Zouros E, Ball AO, Saavedra C, Freeman KR (1994) An Unusual Type of Mitochondrial-<br />

DNA Inheritance in the Blue Mussel Mytilus. Proceedings of the National Aca<strong>de</strong>my of<br />

Sciences of the United States of America 91, 7463-7467.<br />

321


Références Bibliographiques<br />

322


Annexe 1<br />

Annexe 1 : Marquage Recapture <strong>de</strong> séquences<br />

323


Annexe 1<br />

324


Annexe 1<br />

325


Annexe 1<br />

326


Annexe 1<br />

327


Annexe 1<br />

328


Annexe 2<br />

Annexe 2 : logiciel Rabouter Séquence<br />

Manuel d'utilisation du programme RabouterSequences<br />

Réalisé par Sophie Plouviez et Matthieu Bruneaux<br />

Equipe Génétique <strong>de</strong> l'Adaptation en Milieux Extrêmes<br />

<strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong><br />

Pourquoi utiliser RabouterSequences?<br />

Le programme RabouterSequences permet <strong>de</strong> rabouter facilement un grand nombre <strong>de</strong><br />

séquences forward et reverse en quelques secon<strong>de</strong>s à partir d'un site <strong>de</strong> coupure que l'on<br />

définit. Il peut être très utile <strong>pour</strong> rabouter <strong>de</strong>s séquences effectuées sur plusieurs individus<br />

d'un même gène. En effet, avant <strong>de</strong> rabouter les séquences forward et reverse, l'utilisateur<br />

vérifie les chromatogrammes <strong>de</strong> ces séquences. La séquence forward est corrigée jusqu'à<br />

un site <strong>de</strong> coupure déterminé appartenant à la zone <strong>de</strong> recouvrement <strong>de</strong>s <strong>de</strong>ux séquences.<br />

Ce site <strong>de</strong> coupure sera le même <strong>pour</strong> l'ensemble <strong>de</strong>s séquences. La séquence reverse est<br />

ensuite corrigée à partir du site <strong>de</strong> coupure jusqu'à la fin <strong>de</strong> la séquence. Le programme<br />

RabouterSequences permettra <strong>de</strong> rabouter chaque séquence forward à sa séquence reverse<br />

correspondante.<br />

Avantage <strong>de</strong> RabouterSequences?<br />

Vous gar<strong>de</strong>z le contrôle sur ce que vous raboutez. Pas <strong>de</strong> séquences consensus. Le logiciel<br />

ne modifie pas les sites <strong>de</strong> vos séquences.<br />

Pas besoin <strong>de</strong> changer l'ordre <strong>de</strong> vos séquences <strong>pour</strong> que Forward et Reverse soient<br />

homogènes. Le logiciel cherche lui même la séquence correspondante.<br />

Rapi<strong>de</strong> et simple d'utilisation (exemple: 600 séquences raboutées en 25 secon<strong>de</strong>s, alors que<br />

manuellement cela vous prendrez plusieurs heures voire jours).<br />

Comment utiliser RabouterSequences?<br />

Fichier d'entrée:<br />

Le programme nécessite <strong>de</strong>ux fichiers d'entrée au format fasta (.fas) contenant<br />

respectivement toutes les séquences en forward (voir exempleForward.fas) et toutes les<br />

séquences en reverse (voir exempleReverse.fas). Les noms <strong>de</strong>s séquences ne doivent pas<br />

dépasser 50 caractères. Il est inutile <strong>de</strong> tronquer les séquences forward et reverse <strong>de</strong> part et<br />

d'autre du site <strong>de</strong> coupure. Ainsi, si le site <strong>de</strong> coupure est 500pb, les 500 premiers<br />

caractères <strong>de</strong>s séquences du fichier forward seront raboutés aux caractères 501 à la fin <strong>de</strong>s<br />

séquences du fichier reverse.<br />

Attention: le programme reconnait automatiquement le format fasta en .fas. Si votre fichier<br />

est en .fasta, changer le en .fas. Lorsque le nom <strong>de</strong>s fichiers vous est <strong>de</strong>mandé, il est donc<br />

inutile d'inscrire .fas. Dans l'exemple fourni il suffit donc <strong>de</strong> taper:<br />

Nom du fichier Forward? exempleForward<br />

329


Annexe 2<br />

Nom du fichier Reverse? exempleReverse<br />

Paramètres <strong>de</strong>mandés par le programme:<br />

Il est inutile d'indiquer .fas au logiciel <strong>pour</strong> le nom du fichier <strong>de</strong> sortie. Dans l'exemple<br />

fourni:<br />

Nom du fichier <strong>de</strong> sortie? exempleRaboutecoupure110<br />

Le programme compare les premiers caractères <strong>de</strong>s noms <strong>de</strong>s séquences forward aux<br />

premiers caractères <strong>de</strong>s séquences reverse. Si les noms sont i<strong>de</strong>ntiques, il raboute les<br />

séquences <strong>de</strong> part et d'autre du site <strong>de</strong> coupure. Il faut donc lui définir le nombre <strong>de</strong><br />

caractères à comparer entre les <strong>de</strong>ux fichiers. Ainsi, dans l'exemple, les 4 premiers<br />

caractères permettent d'i<strong>de</strong>ntifier la séquence complémentaire (e.g. ind1) et le site <strong>de</strong><br />

coupure a été choisi à 110pb. Dans le fichier <strong>de</strong> sortie, les séquences raboutées auront le<br />

nom ses caractères. Il est possible d'y ajouter jusqu'à 15 caractères <strong>pour</strong> l'ensemble <strong>de</strong>s<br />

séquences. Ainsi, dans l'exemple, nous avons ajouté "Raboute" à toutes les séquences.<br />

Caracteres ajoutes aux noms <strong>de</strong>s sequences raboutees? Raboute<br />

Si vous ne souhaitez pas rajouter <strong>de</strong> caractères aux noms <strong>de</strong> séquences raboutées, tapez<br />

"non".<br />

Attention: Au sein du même fichier, <strong>de</strong>ux séquences ne doivent pas avoir exactement les<br />

mêmes caractères d'i<strong>de</strong>ntification.<br />

Fichier <strong>de</strong> sortie:<br />

Deux fichiers <strong>de</strong> sortie sont founis:<br />

Le premier fichier RabouterJournal (.txt) contient les informations <strong>de</strong> déroulement du<br />

programme: les séquences raboutées (faite), les séquences forward dont aucune séquence<br />

reverse n'a été trouvée (forward seule), les séquences reverse dont aucune séquence<br />

forward n'a été trouvée (reverse seule).<br />

Le second fichier <strong>de</strong> sortie est en fasta (.fas) et contient l'ensemble <strong>de</strong>s séquences<br />

raboutées.<br />

Compatibilité <strong>de</strong> RabouterSequences:<br />

RabouterSequences a été compilé <strong>pour</strong> fonctionner sous Windows. Cependant, son<br />

utilisation sous MacOS ou Linux nécessite <strong>de</strong> recompiler les fichiers sources .c et .h. Ces<br />

fichiers peuvent vous être fournis sur <strong>de</strong>man<strong>de</strong>.<br />

330


Annexe 2<br />

Fichier source .c<br />

// programme <strong>pour</strong> rabouter <strong>de</strong>s sequences<br />

// Sophie et Matthieu<br />

// commence le 26/02/2009<br />

// <strong>de</strong>rniere modification le 01/03/2009<br />

#inclu<strong>de</strong> <br />

#inclu<strong>de</strong> <br />

#inclu<strong>de</strong> <br />

#inclu<strong>de</strong> <br />

#inclu<strong>de</strong> "fonctions_rabouterSequences.h"<br />

#<strong>de</strong>fine LONGUEUR_MAX_SEQUENCES 5000<br />

int main(int argc, char *argv[])<br />

{<br />

// variables<br />

char nomFichierForward[50]={0}; // nom du fichier contenant les<br />

sequences forward<br />

char nomFichierReverse[50]={0}; // nom du fichier contenant les<br />

sequences reverse<br />

char nomFichierSortie[150]={0}; // nom du fichier <strong>de</strong> sortie<br />

char nomFichierJournal[50]="rabouterJournal"; // nom du fichier journal<br />

FILE* fichierForward=NULL; // pointeur vers le fichier<br />

Forward<br />

FILE* fichierReverse=NULL; // pointeur vers le fichier<br />

Reverse<br />

FILE* fichierSortie=NULL;<br />

// pointeur vers le fichier <strong>de</strong> sortie<br />

FILE* fichierJournal=NULL;<br />

// pointeur vers le fichier journal<br />

int nombreSequencesForward=0; // nombre <strong>de</strong> sequences du fichier<br />

Forward<br />

int nombreSequencesReverse=0; // nombre <strong>de</strong> sequences du fichier<br />

Reverse<br />

char nomRecherche[50]={0};<br />

// nom d'une sequence forward<br />

dont on cherche la reverse<br />

char nomCompare[50]={0};<br />

// nom d'une sequence reverse que l'on<br />

compare avec celui <strong>de</strong> la forward<br />

int nomTrouve=0;<br />

// booleen : nom Reverse trouve<br />

correspondant a un Forward<br />

char sequenceForward[LONGUEUR_MAX_SEQUENCES]={0}; //<br />

sequence forward<br />

char sequenceReverse[LONGUEUR_MAX_SEQUENCES]={0}; // sequence<br />

reverse<br />

int longueurCoupeForward=0;<br />

// longueur <strong>de</strong> sequence forward<br />

a conserver<br />

int longueurCoupeReverse=0;<br />

// i<strong>de</strong>m reverse<br />

char ajoutNom[15]={0};<br />

//caracteres ajoutes aux noms <strong>de</strong>s<br />

sequences raboutees<br />

331


Annexe 2<br />

int caracteresNomSequencesCompares=0; //nombre <strong>de</strong> caractères compares <strong>de</strong>s<br />

noms <strong>de</strong> sequences<br />

// <strong>de</strong>man<strong>de</strong>r les noms <strong>de</strong> fichiers d'entree et autres<br />

printf("\nRabouterSequences - programme realise par Sophie Plouviez et Matthieu<br />

Bruneaux");<br />

printf("\nEquipe GAME (Genetique <strong>de</strong> l'Adaptation en Milieux Extremes)");<br />

printf("\n<strong>Station</strong> <strong>Biologique</strong> <strong>de</strong> <strong>Roscoff</strong>");<br />

printf("\n\nNom du fichier Forward ? ");<br />

scanf("%s",nomFichierForward);<br />

printf("Nom du fichier Reverse ? ");<br />

scanf("%s",nomFichierReverse);<br />

printf("Nom du fichier <strong>de</strong> sortie ? ");<br />

scanf("%s",nomFichierSortie);<br />

printf("Longueur <strong>de</strong> coupe Forward ? ");<br />

scanf("%ld",&longueurCoupeForward);<br />

printf("Caracteres ajoutes aux noms <strong>de</strong>s sequences raboutees \n(ne rien ajouter:<br />

taper non)? ");<br />

scanf("%s",ajoutNom);<br />

longueurCoupeReverse=longueurCoupeForward+1;<br />

printf("Combien <strong>de</strong> caracteres compares <strong>pour</strong> les noms <strong>de</strong> sequences ? ");<br />

scanf("%ld",&caracteresNomSequencesCompares);<br />

// preparation du nom <strong>de</strong>s fichiers<br />

strcat(nomFichierJournal,"_");<br />

strcat(nomFichierJournal,nomFichierSortie);<br />

strcat(nomFichierJournal,".txt");<br />

strcat(nomFichierForward,".fas");<br />

strcat(nomFichierReverse,".fas");<br />

strcat(nomFichierSortie,".fas");<br />

//variables sequences<br />

char sequenceCoupeForward[LONGUEUR_MAX_SEQUENCES]={0};<br />

char sequenceCoupeReverse[LONGUEUR_MAX_SEQUENCES]={0};<br />

// verification <strong>de</strong>s noms <strong>de</strong> fichiers<br />

printf("\nLes fichiers sont :");<br />

printf("\n\tsequences Forward : %s",nomFichierForward);<br />

printf("\n\tsequences Reverse : %s\n",nomFichierReverse);<br />

// ouverture <strong>de</strong>s fichiers contenant les sequences et du fichier journal<br />

fichierForward=fopen(nomFichierForward,"r");<br />

if (fichierForward==NULL) {<br />

printf("\nErreur : impossible d'ouvrir le fichier %s\n",nomFichierForward);<br />

exit(EXIT_FAILURE);<br />

}<br />

fichierReverse=fopen(nomFichierReverse,"r");<br />

if (fichierReverse==NULL) {<br />

printf("\nErreur : impossible d'ouvrir le fichier %s\n",nomFichierReverse);<br />

332


Annexe 2<br />

exit(EXIT_FAILURE);<br />

}<br />

fichierSortie=fopen(nomFichierSortie,"w+");<br />

if (fichierSortie==NULL) {<br />

printf("\nErreur : impossible d'ouvrir le fichier %s\n",nomFichierSortie);<br />

exit(EXIT_FAILURE);<br />

}<br />

fichierJournal=fopen(nomFichierJournal,"w+");<br />

if (fichierJournal==NULL) {<br />

printf("\nErreur : impossible d'ouvrir le fichier %s\n",nomFichierJournal);<br />

exit(EXIT_FAILURE);<br />

}<br />

// journal<br />

time_t heure;<br />

struct tm* heureConvertie;<br />

time(&heure);<br />

heureConvertie=localtime(&heure);<br />

fprintf(fichierJournal,"rabouterSequences - fichier journal\n");<br />

fprintf(fichierJournal,"\n - programme compile le %s a<br />

%s",__DATE__,__TIME__);<br />

fprintf(fichierJournal,"\n\n\n\n\tDebut du programme :<br />

%s",asctime(heureConvertie));<br />

fprintf(fichierJournal,"\n\nFichiers :");<br />

fprintf(fichierJournal,"\n----------");<br />

fprintf(fichierJournal,"\n Forward : %s",nomFichierForward);<br />

fprintf(fichierJournal,"\n Reverse : %s",nomFichierReverse);<br />

fprintf(fichierJournal,"\n sortie : %s\n",nomFichierSortie);<br />

// compter le nombre <strong>de</strong> sequences dans chaque fichier<br />

nombreSequencesForward=compterSequences(fichierForward);<br />

printf("\nLe fichier %s contient %ld<br />

sequences",nomFichierForward,nombreSequencesForward);<br />

nombreSequencesReverse=compterSequences(fichierReverse);<br />

printf("\nLe fichier %s contient %ld<br />

sequences\n",nomFichierReverse,nombreSequencesReverse);<br />

// journal<br />

fprintf(fichierJournal,"\n Forward : %ld sequences",nombreSequencesForward);<br />

fprintf(fichierJournal,"\n Reverse : %ld sequences\n",nombreSequencesReverse);<br />

//<br />

fprintf(fichierJournal,"\n\n\nRaboutage <strong>de</strong>s sequences :");<br />

fprintf(fichierJournal,"\n-------------------------");<br />

// remise <strong>de</strong>s curseurs au <strong>de</strong>but <strong>de</strong>s fichiers<br />

rewind(fichierForward);<br />

rewind(fichierReverse);<br />

// lecture d'un nom forward et recherche <strong>de</strong> la sequence reverse correspondante<br />

333


Annexe 2<br />

int continuerLectureForward=1;<br />

int continuerLectureReverse=1;<br />

int compteurSequencesLuesForward=0;<br />

int compteurSequencesTrouvees=0;<br />

while<br />

(continuerLectureForward&&compteurSequencesLuesForward


Annexe 2<br />

// fin <strong>de</strong> la boucle <strong>pour</strong> une sequence forward / retour au <strong>de</strong>but<br />

}<br />

// journal<br />

fprintf(fichierJournal,"\n\nTotal : %ld sequences<br />

raboutees\n",compteurSequencesTrouvees);<br />

// sequences forward sans reverse<br />

rewind(fichierForward);<br />

rewind(fichierReverse);<br />

int compteurForwardSeules=0;<br />

continuerLectureForward=1;<br />

continuerLectureReverse=1;<br />

compteurSequencesLuesForward=0;<br />

fprintf(fichierJournal,"\n\n\nSequences Forward seules :");<br />

fprintf(fichierJournal,"\n--------------------------");<br />

while<br />

(continuerLectureForward&&compteurSequencesLuesForward


Annexe 2<br />

sequenceCoupeForward[i]=0;<br />

sequenceCoupeReverse[i]=0;<br />

}<br />

// fin <strong>de</strong> la boucle <strong>pour</strong> une sequence forward / retour au <strong>de</strong>but<br />

}<br />

// journal<br />

fprintf(fichierJournal,"\n\nTotal : %ld sequences Forward<br />

seules\n",compteurForwardSeules);<br />

// sequences reverse sans forward<br />

rewind(fichierForward);<br />

rewind(fichierReverse);<br />

int compteurReverseSeules=0;<br />

continuerLectureForward=1;<br />

continuerLectureReverse=1;<br />

int compteurSequencesLuesReverse=0;<br />

fprintf(fichierJournal,"\n\n\nSequences Reverse seules :");<br />

fprintf(fichierJournal,"\n--------------------------");<br />

while<br />

(continuerLectureReverse&&compteurSequencesLuesReverse


Annexe 2<br />

int i=0;<br />

// vi<strong>de</strong>r les variables temporaires<br />

for (i=0; i


Annexe 2<br />

Fichier source fonction .h<br />

#<strong>de</strong>fine NOMBRE_MAX_CHAR 80<br />

par ligne <strong>de</strong> fichier contenant les sequences<br />

// nombres <strong>de</strong> caracteres maximal lus<br />

int compterSequences(FILE* fichierOuvert) {<br />

int nombreSequences=0;<br />

char ligneLue[NOMBRE_MAX_CHAR];<br />

char* reference=">";<br />

// lecture <strong>de</strong>s lignes<br />

while(fgets(ligneLue,NOMBRE_MAX_CHAR,fichierOuvert)!=NULL) {<br />

if (ligneLue[0]==*reference) {<br />

// on teste si la ligne contient un<br />

nom <strong>de</strong> sequence ('>' en format FASTA)<br />

nombreSequences=nombreSequences+1;<br />

}<br />

}<br />

// renvoie le nombre <strong>de</strong> sequences<br />

return(nombreSequences);<br />

}<br />

void extractionChaine(char* sousChaine, char* chaine, int <strong>de</strong>part, int fin) {<br />

int j=0;<br />

// copie <strong>de</strong>s caracteres un a un (boucle)<br />

for (j=<strong>de</strong>part; j


Annexe 2<br />

extractionChaine(nomSequence,ligneLue,1,longueurComparee); //<br />

extrait le nom <strong>de</strong> la sequence<br />

}<br />

} while ((testLigne!=NULL)&&(nomTrouve==0));<br />

on<br />

if (testLigne==NULL) {succes=0;}<br />

// recul du curseur <strong>pour</strong> etre en <strong>de</strong>but <strong>de</strong> sequence<br />

fseek(fichierOuvert,-1,SEEK_CUR);<br />

// recuperation <strong>de</strong> la sequence<br />

int i=0;<br />

int continuer=1;<br />

while (continuer) {<br />

caractere=fgetc(fichierOuvert);<br />

if ((caractere!=EOF)&&(caractere!=10)&&(caractere!=62)) {<br />

sequenceForward[i]=caractere;<br />

i++;<br />

}<br />

else if ((caractere==EOF)||(caractere==62)) {continuer=0;}<br />

}<br />

fseek(fichierOuvert,-1,SEEK_CUR);<br />

return (succes);<br />

}<br />

int extraireNomSequenceSeul(char* nomSequence, FILE* fichierOuvert, int<br />

longueurComparee) {<br />

int succes=1;<br />

int nombreSequences=0;<br />

char ligneLue[NOMBRE_MAX_CHAR];<br />

char* reference=">";<br />

int nomTrouve=0;<br />

int caractere=0;<br />

char* testLigne=NULL;<br />

// recuperation du nom<br />

do {<br />

testLigne=fgets(ligneLue,NOMBRE_MAX_CHAR,fichierOuvert);<br />

if (ligneLue[0]==*reference) {<br />

// on teste si la ligne contient un nom<br />

<strong>de</strong> sequence ('>' en format FASTA)<br />

nomTrouve=1;<br />

extractionChaine(nomSequence,ligneLue,1,longueurComparee); //<br />

extrait le nom <strong>de</strong> la sequence<br />

}<br />

} while ((testLigne!=NULL)&&(nomTrouve==0));<br />

if (testLigne==NULL) {succes=0;}<br />

// recul du curseur <strong>pour</strong> etre en <strong>de</strong>but <strong>de</strong> sequence<br />

on<br />

339


Annexe 2<br />

fseek(fichierOuvert,-1,SEEK_CUR);<br />

return (succes);<br />

}<br />

340


Annexe 3<br />

Annexe 3 : Conditions <strong>de</strong> PCR<br />

• Alvinella pompejana<br />

Amorces utilisées <strong>pour</strong> l’amplification <strong>de</strong>s gènes nucléaires d’A. pompejana<br />

---- signifie que 4 pb ont été ajoutées <strong>pour</strong> la métho<strong>de</strong> <strong>de</strong> Marquage Recapture <strong>de</strong><br />

séquences<br />

Locus Sens <strong>de</strong> Amorce 5’-3’<br />

l’amorce<br />

PGM Sens ----CCCAGGTGGTCCAAATGC<br />

Antisens ----GTACAGATGTCAGCCTTCAGGTC<br />

SAHH Sens ----CTTGTCAACCTCGGCTGTG<br />

Antisens ----CTGCCTGCTCGTCTGTTAG<br />

GlobX Sens ----AGTTGACTGAAGAACGCAGAGAAGCGGT<br />

Antisens ----CGCGAAGATCCTTGAAGTACTCTTGGTA<br />

Conditions <strong>de</strong> mix <strong>de</strong> PCR <strong>pour</strong> un puit<br />

PGM SAHH GlobX<br />

BSA 2% 0,5 0 0<br />

Tampon Taq 10X 2,5 2,5 2,5<br />

MgCl2 25mM 2 2 2<br />

dNTPs 2,5 mM chacun 2,5 0,5 0,5<br />

Amorce Sens 10µM 1 1,2 1,2<br />

Amorce Antisens 10µM 1 1,2 1,2<br />

Taq 5u/µl 0,1 0,1 0,1<br />

ADN 2,5 5 5<br />

Eau ultrapure 12,9 12,5 12,5<br />

Volume final 25 25 25<br />

Conditions d’amplification <strong>de</strong>s gènes nucléaires.<br />

Tm PGM = 57°C, Tm SAHH = 57°C, Tm GlobX = 55°C<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 3 min 1<br />

94 30 s<br />

Tm<br />

20 s<br />

40<br />

72 2min 30 s<br />

72 10 min 1<br />

341


Annexe 3<br />

• Bathymodiolus thermophilus<br />

Amorces utilisées <strong>pour</strong> l’amplification <strong>de</strong>s gènes nucléaires <strong>de</strong> B. thermophilus<br />

---- signifie que 4 pb ont été ajoutées <strong>pour</strong> la métho<strong>de</strong> <strong>de</strong> Marquage Recapture <strong>de</strong><br />

séquences<br />

Locus Sens <strong>de</strong> Amorce 5’-3’<br />

l’amorce<br />

SAHH Sens ----TAAATCTTGGTTGTGCTCATGGTCATCCA<br />

Antisens ----TTTGAATGGTCCTTCTTTAGGTAGAC<br />

Lysozyme Sens ----GTTTCCCCAAAATGTATGAGCTGT<br />

Antisens ----TAATCTCCGCCTGGACTACCACAATC<br />

Sulfo 1 et 2 Sens ----TCTTTAAAGTCAGGATCACATTGG<br />

Antisens ----TAAGGCAAAGTGGAACAACGAGACCGC<br />

Conditions <strong>de</strong> mix <strong>de</strong> PCR <strong>pour</strong> un puit<br />

SAHH Lysozyme Sulfo1 et 2<br />

BSA 2% 0,5 0,5 0,5<br />

Tampon Taq 10X 2 2 2<br />

MgCl2 25mM 1,6 1,6 1,6<br />

dNTPs 2,5 mM chacun 2 2 2,4<br />

Amorce Sens 10µM 0,8 0,8 0,8<br />

Amorce Antisens 10µM 0,8 0,8 0,8<br />

Taq 5u/µl 0,1 0,1 0,1<br />

ADN 2 2 2<br />

Eau ultrapure 10,2 10,2 9,8<br />

Volume final 20 20 20<br />

Conditions d’amplification <strong>de</strong>s gènes nucléaires<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 5 min 1<br />

94 35 s<br />

Tm 60°C<br />

1 min<br />

35<br />

72 2 min<br />

72 10 min 1<br />

342


Annexe 3<br />

• Lepetodrilus elevatus<br />

Amorces utilisées <strong>pour</strong> l’amplification <strong>de</strong>s gènes nucléaires et l’amplification et le<br />

séquençage <strong>de</strong>s clones mitochondriaux <strong>de</strong> L. elevatus<br />

---- signifie que 4 pb ont été ajoutées <strong>pour</strong> la métho<strong>de</strong> <strong>de</strong> Marquage Recapture <strong>de</strong><br />

séquences<br />

Locus Sens <strong>de</strong> Amorce 5’-3’<br />

l’amorce<br />

Lysozyme Sens ----GAATGAATCGCTCGTAAAATCCC<br />

Antisens ----CGCAGGTAGGGGCACGACC<br />

VDG3 Sens ----CTTCTTCTGTAAGTATCCCYGTC<br />

Antisens ----AAGCCGAACTTCGTCACAA<br />

Textilinine Sens ----GGGGATCCATGTTGAAGTTTGTCGTCTGCTCCTTGT<br />

Antisens ----ATCCCACTCCTGCGCACTTGTTCTCGCAGTCG<br />

Plasmi<strong>de</strong> Sens SP6 CATTTAGGTGACACTATAG<br />

Antisens T7 GTAATACGACTCACTATA<br />

Conditions <strong>de</strong> mix <strong>de</strong> PCR <strong>pour</strong> un puit<br />

Lysozyme VDG3 Textilinine SP6/T7<br />

BSA 2% 0 1 1 0<br />

Tampon Taq 10X 2,5 2,5 2,5 1,25<br />

MgCl2 25mM 2 2 2 1<br />

dNTPs 2,5 mM chacun 0,5 0,7 0,7 0,35<br />

Amorce Sens 10µM 1,2 1,5 1,5 0,75<br />

Amorce Antisens 10µM 1,2 1,5 1,5 0,75<br />

Taq 5u/µl 0,1 0,1 0,1 0,05<br />

ADN 5 5 5 5<br />

Eau ultrapure 12,5 10,7 10,7 3,35<br />

Volume final 25 25 25 12,5<br />

343


Annexe 3<br />

Conditions d’amplification du gène Lysozyme<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 3 min 1<br />

94 30s<br />

TD 66-56°C<br />

20s<br />

10<br />

72 2min 30s<br />

94 30 s<br />

Tm 56°C<br />

20 s<br />

30<br />

72 2min 30s<br />

72 10 min 1<br />

Conditions d’amplification du gène VDG3<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 3 min 1<br />

94 45 s<br />

Tm 57°C<br />

30 s<br />

40<br />

72 1min 15s<br />

72 10 min 1<br />

Conditions d’amplification du gène Textilinine<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 3 min 1<br />

94 45 s<br />

Tm 60°C<br />

30 s<br />

40<br />

72 1min 15s<br />

72 10 min 1<br />

Conditions d’amplification <strong>de</strong>s clones <strong>de</strong> mtCOI<br />

Température °C Temps Nombre <strong>de</strong> cycles<br />

94 3 min 1<br />

94 1 min<br />

Tm 55°C<br />

1 min<br />

30<br />

72 1min 15s<br />

72 10 min 1<br />

344


Annexe 4<br />

Annexe 4 : Co<strong>de</strong> génétique mitochondrial <strong>de</strong>s Alvinellidae<br />

La traduction <strong>de</strong>s séquences d’A. pompejana en utilisant le co<strong>de</strong> mitochondrial <strong>de</strong>s<br />

Plathelminthes « flatworm » met en évi<strong>de</strong>nce six mutations non synonymes dans le<br />

polymorphisme du gène <strong>de</strong> la Cytochrome c Oxydase I (COI). Ces mutations ne sont pas<br />

fixées entre les lignées nord et sud, ne correspon<strong>de</strong>nt pas à <strong>de</strong>s singletons mais sont plutôt<br />

en équi-fréquence dans le jeu <strong>de</strong> séquences analysées (mutations en fréquence<br />

intermédiaire). Elles correspon<strong>de</strong>nt toutes à une mutation entre l’aci<strong>de</strong> aminé méthionine<br />

(ATG) et un unique codon <strong>de</strong> l’isoleucine (ATA, Fig. A.1). Cette observation suggère que<br />

le codon ATA <strong>pour</strong>rait co<strong>de</strong>r en réalité <strong>pour</strong> une méthionine et non <strong>pour</strong> une isoleucine :<br />

une situation déjà rencontrée dans le co<strong>de</strong> génétique <strong>de</strong> la mitochondrie <strong>de</strong> la drosophile et<br />

<strong>de</strong>s Chordés (e.g. Wolstenholme & Clary 1985). En utilisant le co<strong>de</strong> génétique <strong>de</strong> la<br />

drosophile (plus proche phylogénétiquement), aucune mutation non synonyme n’est<br />

présente, ni dans le polymorphisme d’A. pompejana, ni dans sa divergence avec A.<br />

caudata, indiquant une forte sélection purifiante sur le gène mtCOI. Avec le co<strong>de</strong><br />

génétique <strong>de</strong>s Plathelminthes, le test <strong>de</strong> McDonald Kreitman réalisé entre A. pompejana et<br />

A. caudata est significatif (p = 0,04), avec 6 mutations non synonymes dans le<br />

polymorphisme d’A. pompejana mais aucune dans la divergence fixée entre ces espèces,<br />

confirmant que le biais observé est dû à l’usage d’un mauvais co<strong>de</strong> génétique.<br />

345


Annexe 4<br />

10 20 30 40 50 60 70 80 90 100<br />

....|....| ....|....| ....|....| ....|....| ....|....| ....|....| ....|....| ....|....| ....|....| ....|....|<br />

Ap11 DDQLYNTIIT AHGLLMLFFF VTPIFIGGFA NWLLPLMIGA PDMAFPRVNN LGFWLLPPAL TLLIASAAVE KGVGTGWTIY PPLAGNMAHA GPSVDLAIFS<br />

Ap45 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap63 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap82 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap98 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap107 .......... .......... .......... ......I... .......... .......... .......... .......... .......... ..........<br />

Ap134 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap281 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap292 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap362 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap500 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap514 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap525 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap530 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap542 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap553 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap564 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap9N4 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ap9N6 .......... .......... .......... .......... .......... .......... .......... .......... ......I... ..........<br />

Ap9N12 .......... .......... .......... .......... .......... .......... .......... .......... ......I... ..........<br />

Ac18 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ac31 .......... .......... .......... .......... ..I....... .......... .......... .......... .......... ..........<br />

Ac33 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Ac38 .......... .......... .......... .......... .......... .......... .......... .......... .......... ..........<br />

Pp1 .......... .....K.... I........G .......L.. ..I....M.. .R...N.... ....T..... ..I....... .......... ..........<br />

Pp2 .......... .......... I........G .......L.. ..I....M.. .R...N.... ....T..... ..M....... .L........ ..........<br />

Pp3 .......... .......... I........G .......L.. ..I....M.. .R...N.... ....T..... ..M....... .......... ..........<br />

Pp4 .......... .......... I........G .......L.. ..I....M.. .R...N.... ....T..... ..M....... .......... ..........<br />

Pp5 .......... .......... I........G .......L.. ..I....M.. .R...N.... ....T..... ..M....... .......... ..........<br />

Pk1 R......VV. ..AF.II..L .I.V.....G ...V..ILA. .......I.. MR........ I..LS..... ........V. ...SR.L... ..........<br />

Pk2 R......VV. ..AF.II..L .I.V.....G ...V..ILA. .......I.. MR........ I..LS..... ........V. ...SR.L... ..........<br />

Pk3 R......VV. ..AF.II..L .I.V.....G ...V..ILA. .......I.. IS........ I..LS..... ........V. ...SR.L... ..........<br />

Pk4 R......VV. ..AF.II..L .I.V.....G ...V..ILA. .......I.. IS........ I..LS..... ........V. ...SR.L... ..........<br />

Pk5 R......VV. ..AF.II..L .I.V.....G ...V..ILA. .......I.. IS........ I..LS..... ........V. ...SR.L... ..........<br />

110 120 130 140 150<br />

....|....| ....|....| ....|....| ....|....| ....|....| ....|.<br />

Ap11 LHLAGVSSIL GSINYITTVM NMRPKGMNFE SVPLFVWAVK ITAILLLLSL PVFAGA<br />

Ap45 .......... .......... .......... .......... .......... ......<br />

Ap63 .......... .......... .......... .......... .......... ......<br />

Ap82 .......... .......... .......... .......... .......... ......<br />

Ap98 .......... .......... .......... .......... .......... ......<br />

Ap107 .......... .......... .......... .......... .......... ......<br />

Ap134 .......... .......... .......... .......... .......... ......<br />

Ap281 .......... .......... .......... .......... .......... ......<br />

Ap292 .......... .......... .......... .......... .......... ......<br />

Ap362 .......... .......... .......... .......... .......... ......<br />

Ap500 .......... .........I .......... .......... .......... ......<br />

Ap514 .......... .......... .I........ .......... .......... ......<br />

Ap525 .......... .......... .I........ .......... .......... ......<br />

Ap530 .......... .......... .I........ .......... .......... ......<br />

Ap542 .......... .......... .I........ .......... .......... ......<br />

Ap553 .......... .......... .I........ .......... .......... ......<br />

Ap564 .......... .......... .I........ .......... .......... ......<br />

Ap9N4 .......... .......... .......... .......... .......... ......<br />

Ap9N6 .......... .......... .I........ .......... .......... ......<br />

Ap9N12 .......... .......... .I........ .......... .......... ......<br />

Ac18 .......... .........I .......... .......... .......... ......<br />

Ac31 .......... .........I .......... .......... .......... ......<br />

Ac33 .......... .........I .......... .......... .......... ......<br />

Ac38 .......... .........I .......... .......... .......... ......<br />

Pp1 .....A.A.. ....F..... ......I... .......S.. .......... ......<br />

Pp2 .....A.A.. ....F..... ......I... .......S.. .......... ......<br />

Pp3 .....A.A.. ....F..... ......I... .......S.. .......... ......<br />

Pp4 .....A.A.. ....F..... ......I... .......S.. .......... ......<br />

Pp5 .....A.A.. ....F..... ......I... .......S.. .......... ......<br />

Pk1 .....I.... .A..F....I ...S..LRL. R......... ...V...... ..L...<br />

Pk2 .....I.... .A..F....I ...S..LRL. R......... ...V...... ..L...<br />

Pk3 .....I.... .A..F....I ...S..LRL. R......... .......... ..L...<br />

Pk4 .....I.... .A..F....I ...S..LRL. RI........ .......... ..L...<br />

Pk5 .....I.... .A..F....I ...S..LRL. R......... .......... ..L...<br />

Fig. A.1 Alignement <strong>de</strong> séquences <strong>de</strong> mtCOI d’A. pompejana (Ap) et A. caudata (Ac)<br />

traduites en aci<strong>de</strong>s aminés en utilisant le co<strong>de</strong> génétique <strong>de</strong>s Plathelminthes.<br />

346


Annexe 4<br />

Nous avons donc cherché à vérifier cette hypothèse en comparant ces séquences<br />

d’A. pompejana à <strong>de</strong>s séquences <strong>de</strong> mtCOI d’autres Alvinellidae (A. caudata, Paralvinella<br />

pandorae) et un polychète côtier d’une famille proche (Pectinaria koreni). Nous avons<br />

effectué un arbre phylogénétique en utilisant un sous-échantillon <strong>de</strong> 20 séquences d’A.<br />

pompejana, selon la métho<strong>de</strong> <strong>de</strong>s plus proches voisins : sur les séquences en aci<strong>de</strong>s aminés<br />

traduites selon les <strong>de</strong>ux co<strong>de</strong>s génétiques (Fig. A.2). Avec le co<strong>de</strong> génétique <strong>de</strong>s<br />

Plathelminthes, un individu A. pompejana s’avère i<strong>de</strong>ntique en aci<strong>de</strong>s aminés à plusieurs<br />

A. caudata. Le co<strong>de</strong> génétique mitochondrial <strong>de</strong> Drosophila montre, quant à lui, que<br />

l’ensemble <strong>de</strong>s Alvinella présente la même séquence d’aci<strong>de</strong>s aminés (râteau), suggérant<br />

que le co<strong>de</strong> génétique mitochondrial <strong>de</strong> Drosophila correspondrait mieux à celui <strong>de</strong>s<br />

Alvinella.<br />

Fig. A.2 Arbres phylogénétiques du gène mtCOI <strong>de</strong> séquences en aci<strong>de</strong>s aminés avec le<br />

co<strong>de</strong> <strong>de</strong>s Plathelminthes (A) et <strong>de</strong> Drosophila (B).<br />

347


Annexe 4<br />

348


Annexe 5<br />

Annexe 5 : Article Relation femelles/juvéniles chez<br />

Branchipolynoe seepensis<br />

349


Annexe 5<br />

350


Annexe 5<br />

351


Annexe 5<br />

352


Annexe 5<br />

353


Annexe 5<br />

354


Annexe 5<br />

355


Annexe 5<br />

356


Annexe 5<br />

357


358


359


Titre <strong>de</strong> thèse : Phylogéographie comparée <strong>de</strong>s espèces hydrothermales <strong>de</strong> la dorsale<br />

du Pacifique oriental<br />

L’histoire démographique <strong>de</strong>s espèces hydrothermales profon<strong>de</strong>s et leurs échanges<br />

génétiques le long <strong>de</strong> la dorsale du Pacifique oriental ont été analysés en combinant <strong>de</strong>s<br />

approches classiques d’analyses <strong>de</strong> fréquences alléliques entre populations avec <strong>de</strong>s<br />

métho<strong>de</strong>s d’analyse plus récentes basées sur la théorie <strong>de</strong> la coalescence. Une approche <strong>de</strong><br />

phylogéographie comparée sur sept espèces (un bivalve, trois gastéropo<strong>de</strong>s, trois<br />

polychètes), réalisée sur le gène mitochondrial Cytochrome Oxydase I, a permis <strong>de</strong> révéler<br />

une barrière aux flux <strong>de</strong> gènes, commune à toutes les espèces entre le Nord et le Sud EPR<br />

datant d’environ 1,3 Ma et probablement liée à la formation <strong>de</strong> failles transformantes entre<br />

0°-7°25°N. Cette séparation <strong>de</strong> faune s’est accompagnée d’une expansion démographique<br />

généralisée plus récente (< 0,5 Ma) au Sud. Une approche multi-locus a ensuite été<br />

effectuée sur trois taxons cibles aux traits d’histoire <strong>de</strong> vie contrastés : le polychète A.<br />

pompejana, le bivalve B. thermophilus et le gastéropo<strong>de</strong> L. elevatus (comprenant <strong>de</strong>ux<br />

espèces cryptiques). Cette approche a permis <strong>de</strong> confirmer les résultats obtenus par<br />

l’approche multi-spécifique. L’hybridation entre lignées divergentes, détectée chez ces<br />

trois taxons, suggère une semi-perméabilité <strong>de</strong> la barrière aux flux <strong>de</strong> gènes. L’étendue <strong>de</strong><br />

la zone <strong>de</strong> contact secondaire dépend du taxon considéré selon ses traits d’histoire <strong>de</strong> vie,<br />

et <strong>de</strong>s forces <strong>de</strong> sélection agissant au locus considéré. Les patrons <strong>de</strong> distributions<br />

phylogéographique et biogéographique mises en évi<strong>de</strong>nce sur les espèces et les<br />

communautés s’accor<strong>de</strong>nt <strong>pour</strong> séparer la dorsale du Pacifique oriental en <strong>de</strong>ux provinces<br />

biogéographiques différentes <strong>de</strong> part et d’autre <strong>de</strong> l’Equateur.<br />

Mots clés : phylogéographie, vicariance, allopatrie, expansion démographique,<br />

hybridation, introgression, sélection, milieu hydrothermal<br />

PhD title: Comparative phylogeography of <strong>de</strong>ep-sea hydrothermal vents species along<br />

the East Pacific Rise<br />

Demographic history and past gene flow between populations of <strong>de</strong>ep-sea<br />

hydrothermal vent species along the East Pacific Rise (EPR) were assessed using a<br />

phylogeographic approach combining a classical allele-frequency-based analysis and more<br />

sophisticated coalescence-based methods. A comparative phylogeographic COI-based<br />

analysis of seven species (one bivalve, three gastropods, three polychaetes) showed the<br />

occurrence of a vicariant event associated with the raise of a physical barrier between the<br />

Northern and Southern EPR, 1,3 Mya probably due to the formation of transform faults in<br />

the equatorial region (0°-7°25S). This fauna separation was then followed by a more recent<br />

(< 0.5 Myr) population expansion in the South for all species. A multilocus approach was<br />

then performed on three targeted taxa with different life trait histories (the polychaete: A.<br />

pompejana, the bivalve: B. thermophilus, and two cryptic species of the gastropod L.<br />

elevatus). This study confirmed results obtained from the previous multi-species approach.<br />

Hybridization between divergent lineages from each of the three taxa suggested that the<br />

barrier is semi-permeable. The range of the secondary contact zone was however variable<br />

<strong>de</strong>pending on life-history traits of each species and selective processes acting at a given<br />

gene. Phylogeographic and biogeographical patterns <strong>de</strong>rived from species and assemblages<br />

are in agreement with the hypothesis of two biogeographical EPR provinces across the<br />

Equator.<br />

Keywords: phylogeography, vicariance, allopatry, <strong>de</strong>mographic expansion,<br />

hybridization, introgression, selection, <strong>de</strong>ep-sea hydrothermal vents<br />

360

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!